This question has been asked a couple of times and I tried this, this, this etc, but couldn’t solve my problem. I have just 4 images of approximately 1MB which I’m trying to fit into a deep learning model. However, I keep getting the resource exhaust error. The error occurs in the dense layer but I need all the layers for my analysis. Please can someone help with a workround?

Working with 16.0 GB (15.6 GB usable) computer

def discriminator(shape):

    # Adding the input layer and the first hidden layer
    ip = Input(shape = (165, 209, 3))
    
    model = Conv2D(64, kernel_size=3, padding='same', strides=1)(ip)
    model = BatchNormalization(momentum=0.5)(model)
    
    model = layer(disc_model, 64, 3, 1)
    model = layer(disc_model, 128, 3, 1)
    model = layer(disc_model, 128, 3, 2)
    model = layer(disc_model, 256, 3, 2)
    model = layer(disc_model, 256, 3, 2)
    model = layer(disc_model, 512, 3, 2)
    model = layer(disc_model, 512, 3, 2)
    
    model = Flatten()(model)
    model = Dense(1024)(model)
    model = LeakyReLU(alpha = 0.2)(model)
    model = Dense(1, activation = 'sigmoid')(model)
    
    return Model(inputs = ip, outputs = model)

wv_ip = Input(shape = (115, 209, 3))
mod = discriminator(wv_ip)
mod.summary()

mod.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accurcay', 'mean_square_error'])
mod.summary()

ResourceExhaustedError                    Traceback (most recent call last)
Cell In[33], line 2
      1 wv_ip = Input(shape = (115, 209, 3))
----> 2 mod = discriminator(wv_ip)
      3 mod.summary()
      5 mod.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accurcay', 'mean_square_error'])

Cell In[32], line 61, in discriminator(shape)

---> 61 model = Dense(1024)(model)

ResourceExhaustedError: OOM when allocating tensor with shape[17656320,1024] and type float on /job