Quick question...

[JoshK68]

Hi,

I'm getting this error:

history_101_horses_or_humans_feature_extract = model.fit(train_data,

                                                     epochs=5,
                                                     steps_per_epoch=int(0.01 * len(train_data)),
                                                     validation_data=test_data,
                                                     validation_steps=int(0.01 * len(test_data)),
                                                     callbacks=[create_tensorboard_callback('training_logs',
                                                                                            'efficientnetb0_horses_or_humans_feature_extract'),
                                                                                            model_checkpoint])

ValueError: Expect x to be a non-empty array or dataset.

Any idea what's going on or how I can fix it?
I Googled and used SO, tried several things...still stuck on this probably easy challenge.

Here's my code: https://colab.research.google.com/drive/1NA2mv5wozRrxZ0mm1vPLgOQuMIBqQP6q?usp=sharing

Thank you very much!
Josh

[mrdbourke]

Hey Josh,

ValueError: Expect x to be a non-empty array or dataset.

Looks like your dataset is empty, I'd try checking out where you're creating your data and make sure it contains samples before running model.fit() again.

[JoshK68]

Hey Daniel,

My train_data must contain samples. Basically I just copy and pasted the code from your CNN github this time to make sure that I didn't mistype anything.

Still same error: ValueError: Expect x to be a non-empty array or dataset.

But x is NOT empty. See below.
I've been Googling and using SO and don't know how to fix this. Please help!

from tensorflow.keras import layers
from tensorflow.keras.layers.experimental import preprocessing

Create base model

input_shape = (224, 224, 3)
base_model = tf.keras.applications.EfficientNetB0(include_top=False)
base_model.trainable = False # freeze base model layers

Create Functional model

inputs = layers.Input(shape=input_shape, name="input_layer")

Note: EfficientNetBX models have rescaling built-in but if your model didn't you could have a layer like below

x = preprocessing.Rescaling(1./255)(x)

x = base_model(inputs, training=False) # set base_model to inference mode only
x = layers.GlobalAveragePooling2D(name="pooling_layer")(x) ### Use 2D b/c our shapes are 2 Dimensional (height * Width)
x = layers.Dense(len(class_names))(x) ### want one output neuron per class

Separate activation of output layer so we can output float32 activations

outputs = layers.Activation("sigmoid", dtype=tf.float32, name="sigmoid_float32")(x) #### When do we put x on the inside vs outside (x)
model = tf.keras.Model(inputs, outputs)

Compile the model

model.compile(loss="binary_crossentropy", # Use sparse_categorical_crossentropy when labels are not one-hot
optimizer=tf.keras.optimizers.Adam(),
metrics=["accuracy"])

You can see how x is NOT empty: https://colab.research.google.com/drive/1NA2mv5wozRrxZ0mm1vPLgOQuMIBqQP6q?usp=sharing

[mrdbourke]

Hey Josh,

I checked out the code you shared and it looks like it's running - can you confirm this?

Did you manage to fix your error?

[JoshK68]

Hi Daniel, Yes, it took a few days and a lot of frustration, but I was finally able to fix it. :) You don't know how grateful that I am for your course (which I recently finished). I was in a career that I hated, and wanted to switch to tech. Didn't know what I wanted to do...Nor did I have the confidence to pass a technical assessment (or even apply). Shortly after discovering your course...it was like a treasure that I didn't know existed. I knew what I wanted to do and as I progressed and now create little ML projects for myself, became much more competent. The community that you created is also next level. Absolutely in love with it. Thank you for having such a positive impact on my life. Gratefully, Your friend, Josh

…

On Mon, Sep 20, 2021 at 2:06 AM Daniel Bourke ***@***.***> wrote: Hey Josh, I checked out the code you shared and it looks like it's running - can you confirm this? Did you manage to fix your error? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#198 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AQVR7YSQEDCOW5VDHIAO5N3UC3FPDANCNFSM5EFRWTOQ> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[mrdbourke]

Thank you for the kind words Josh, I really appreciate it!

Keep learning legend.