GAN Discriminator model input shape issue - tensorflow

Question

I am trying to create a simple GAN model but I am getting an input error in Discriminator. Any suggestions or help is welcome
seed(33)
tf.random.set_seed(432)
Generator
Input size (32,) output size (9,)
# Define the generator model
def build_generator():
generator_input = Input(shape=(32,))
x = Dense(16, activation='relu')(generator_input)
x = Dense(9, activation='linear')(x)
generator = Model(generator_input, x)
return generator
Discriminator
The model has 2 embeddings for batsman & bowler and 7 other predictors for a total of 9 predictors
# Define the discriminator model
def build_discriminator():
bowlerIdx_input = Input(shape=(1,), name='bowlerIdx')
batsmanIdx_input = Input(shape=(1,), name='batsmanIdx')
ballNum_input = Input(shape=(1,), name='ballNum')
ballsRemaining_input = Input(shape=(1,), name='ballsRemaining')
runs_input = Input(shape=(1,), name='runs')
runRate_input = Input(shape=(1,), name='runRate')
numWickets_input = Input(shape=(1,), name='numWickets')
runsMomentum_input = Input(shape=(1,), name='runsMomentum')
perfIndex_input = Input(shape=(1,), name='perfIndex')
no_of_unique_batman=len(df1["batsmanIdx"].unique())
print(no_of_unique_batman)
no_of_unique_bowler=len(df1["bowlerIdx"].unique())
print(no_of_unique_bowler)
embedding_size_bat = no_of_unique_batman ** (1/4)
print(embedding_size_bat)
embedding_size_bwl = no_of_unique_bowler ** (1/4)
print(embedding_size_bwl)
# create embedding layer for the categorical predictor
batsmanIdx_embedding = Embedding(input_dim=4742, output_dim=16,input_length=1)(batsmanIdx_input)
print(batsmanIdx_embedding)
batsmanIdx_flatten = Flatten()(batsmanIdx_embedding)
print(batsmanIdx_flatten)
bowlerIdx_embedding = Embedding(input_dim=3492, output_dim=16,input_length=1)(bowlerIdx_input)
bowlerIdx_flatten = Flatten()(bowlerIdx_embedding)
print(bowlerIdx_flatten)
# concatenate all the predictors
discriminator_input = keras.layers.concatenate([batsmanIdx_flatten,bowlerIdx_flatten, ballNum_input, ballsRemaining_input, runs_input, runRate_input, numWickets_input, runsMomentum_input, perfIndex_input])
print(discriminator_input.shape)
# add hidden layers
x = Dense(64, activation='relu')(discriminator_input)
x = Dense(32, activation='relu')(x)
x = Dense(16, activation='relu')(x)
x = Dense(8, activation='relu')(x)
# add output layer
output = Dense(1, activation='sigmoid', name='output')(x)
# create model
discriminator = Model(inputs=[batsmanIdx_input, bowlerIdx_input, ballNum_input, ballsRemaining_input, runs_input, runRate_input, numWickets_input, runsMomentum_input, perfIndex_input], outputs=output)
return discriminator
GAN model
The generator output is fed to discriminator. Concatenate synthetic and real data and run discriminator. Minimize loss
# Define the GAN model
def build_gan(generator, discriminator):
gan_input = Input(shape=(1024,))
x = generator(gan_input)
gan_output = discriminator(x)
gan = Model(gan_input, gan_output)
return gan
# Instantiate the generator, discriminator, and GAN models
generator = build_generator()
discriminator = build_discriminator()
discriminator.summary()
gan = build_gan(generator, discriminator)
# Compile the generator and discriminator models
discriminator.compile(optimizer=Adam(learning_rate=.0002, beta_1=0.5), loss='binary_crossentropy', metrics=['accuracy'])
gan.compile(optimizer=Adam(learning_rate=.0005, beta_1=0.5), loss='binary_crossentropy',metrics=['accuracy'])
# Set the batch size and number of epochs
batch_size = 1024
num_epochs = 20
# Store the losses over time
gen_losses = []
dis_losses = []
gen_acc = []
dis_acc = []
# Train the GAN
for epoch in range(num_epochs):
# Generate synthetic data
synthetic_data = generator.predict(np.random.randn(batch_size, 1024))
# Concatenate synthetic data with real data
real_data = train_dataset1.values
data = np.concatenate((synthetic_data, real_data))
# Create labels for synthetic and real data
labels = np.concatenate((np.zeros(batch_size), np.ones(len(real_data))))
# Train the discriminator on synthetic and real data
d_loss = discriminator.fit([data['batsmanIdx'],data['bowlerIdx'],data['ballNum'],data['ballsRemaining'],data['runs'],
data['runRate'],data['numWickets'],data['runsMomentum'],data['perfIndex']], labels)
#d_loss = discriminator.train_on_batch(data, labels)
dis_losses.append(d_loss[0])
dis_acc.append(d_loss[1])
# Generate random noise for the generator
random_noise = np.random.randn(batch_size, 1024)
# Create labels for the generator (all ones, since we want the generator to fool the discriminator)
# Create labels for the generator (all ones, since we want the generator to fool the discriminator)
generator_labels = np.ones(batch_size)
# Train the generator
g_loss = gan.train_on_batch(random_noise, generator_labels)
gen_losses.append(g_loss[0])
gen_acc.append(g_loss[1])
# Print loss values for each epoch
print(f'Epoch: {epoch+1}, Discriminator Loss: {d_loss}, Generator Loss: {g_loss}')
The output and error I get is
Model: "model_56"
__________________________________________________________________________________________________
Layer (type) Output Shape Param # Connected to
==================================================================================================
batsmanIdx (InputLayer) [(None, 1)] 0 []
bowlerIdx (InputLayer) [(None, 1)] 0 []
embedding_2 (Embedding) (None, 1, 16) 75872 ['batsmanIdx[0][0]']
embedding_3 (Embedding) (None, 1, 16) 55872 ['bowlerIdx[0][0]']
flatten_2 (Flatten) (None, 16) 0 ['embedding_2[0][0]']
flatten_3 (Flatten) (None, 16) 0 ['embedding_3[0][0]']
ballNum (InputLayer) [(None, 1)] 0 []
ballsRemaining (InputLayer) [(None, 1)] 0 []
runs (InputLayer) [(None, 1)] 0 []
runRate (InputLayer) [(None, 1)] 0 []
numWickets (InputLayer) [(None, 1)] 0 []
runsMomentum (InputLayer) [(None, 1)] 0 []
perfIndex (InputLayer) [(None, 1)] 0 []
concatenate_28 (Concatenate) (None, 39) 0 ['flatten_2[0][0]',
'flatten_3[0][0]',
'ballNum[0][0]',
'ballsRemaining[0][0]',
'runs[0][0]',
'runRate[0][0]',
'numWickets[0][0]',
'runsMomentum[0][0]',
'perfIndex[0][0]']
dense_228 (Dense) (None, 64) 2560 ['concatenate_28[0][0]']
dropout_111 (Dropout) (None, 64) 0 ['dense_228[0][0]']
dense_229 (Dense) (None, 32) 2080 ['dropout_111[0][0]']
dropout_112 (Dropout) (None, 32) 0 ['dense_229[0][0]']
dense_230 (Dense) (None, 16) 528 ['dropout_112[0][0]']
dropout_113 (Dropout) (None, 16) 0 ['dense_230[0][0]']
dense_231 (Dense) (None, 8) 136 ['dropout_113[0][0]']
dropout_114 (Dropout) (None, 8) 0 ['dense_231[0][0]']
output (Dense) (None, 1) 9 ['dropout_114[0][0]']
==================================================================================================
Total params: 137,057
Trainable params: 137,057
Non-trainable params: 0
__________________________________________________________________________________________________
WARNING:tensorflow:Model was constructed with shape (None, 1) for
input KerasTensor(type_spec=TensorSpec(shape=(None, 1),
dtype=tf.float32, name='batsmanIdx'), name='batsmanIdx',
description="created by layer 'batsmanIdx'"), but it was called
on an input with incompatible shape (None, 9).
(None, 9)
---------------------------------------------------------------------------
AssertionError Traceback (most recent call last)
<ipython-input-30-8cbe40cd27bc> in <module>
94 print("111")
95 discriminator.summary()
---> 96 gan = build_gan(generator, discriminator)
97
98 # Compile the generator and discriminator models
2 frames
/usr/local/lib/python3.8/dist-packages/keras/engine/functional.py in _run_internal_graph(self, inputs, training, mask)
677 for x in self.outputs:
678 x_id = str(id(x))
--> 679 assert x_id in tensor_dict, "Could not compute output " + str(x)
680 output_tensors.append(tensor_dict[x_id].pop())
681
AssertionError: Exception encountered when calling layer "model_56" (type Functional).
Could not compute output KerasTensor(type_spec=TensorSpec(shape=(None, 1), dtype=tf.float32, name=None), name='output/Sigmoid:0', description="created by layer 'output'")
Call arguments received by layer "model_56" (type Functional):
• inputs=tf.Tensor(shape=(None, 9), dtype=float32)
• training=None
• mask=None
I don't understand what this message means. I am supposed to input all 9 predictors but for some reason it only picks the first
WARNING:tensorflow:Model was constructed with shape (None, 1)
for input KerasTensor(type_spec=TensorSpec(shape=(None, 1),
dtype=tf.float32, name='batsmanIdx'), name='batsmanIdx',
description="created by layer 'batsmanIdx'"), but it was called
on an input with incompatible shape (None, 9).
(. None, 9)
All thoughts, suggestions are welcome