CODEWITHSUNDEEP
tensorflowkeras
Tianyun Ling
Tianyun Ling

Reputation: 1097

efficientnet.tfkeras vs tf.keras.applications.efficientnet

I am trying to use efficientnet to custom train my dataset.

And I find out with all other code/data/config the same. efficientnet.tfkeras.EfficientNetB0 can gives ~90% training/prediction accruacy and tf.keras.applications.efficientnet.EfficientNetB0 only gives ~70% accuracy.

But I guess both should be the same implementation of the efficient net, or I am missing something here?

I am using latest efficientnet and Tensorflow 2.3.0

with strategy.scope():
    model = tf.keras.Sequential([
        efficientnet.tfkeras.EfficientNetB0( #tf.keras.applications.efficientnet.EfficientNetB0
            input_shape=(IMAGE_SIZE, IMAGE_SIZE, 3),
            weights='imagenet',
            include_top=False
        ),
        L.GlobalAveragePooling2D(),
        L.Dense(1, activation='sigmoid')
    ])

model.compile(
    optimizer='adam',
    loss='binary_crossentropy',
    metrics=['binary_crossentropy']
)
model.summary()

Upvotes: 3

Views: 2246

Answers (1)

Mark wijkhuizen
Mark wijkhuizen

Reputation: 393

I did run into the same problem for EfficientNetB4 and did encounter the following:

  1. The number of total parameters are not equal. The trainable parameters are equal, but the non-trainable parameters aren't. The efficientnet.tfkeras has 7 fewer non-trainable parameters than the tf.keras.applications model.

  2. The number of layers are not equal, the efficientnet.tfkeras has fewer layers than tf.keras.application model.

  3. The different layers are at the very beginning, the most noteworthy are the normalization and rescaling layers, which are in the tf.keras.applications model, but not in the efficientnet.tfkeras model. You can observe this yourself using the model.summary() method.

  4. When applying this layer, by using model.layers[i](array), it turn out these layers do rescale the image by dividing it by 255 and applying normalization according to:

(input_image - IMAGENET_MEAN) / square_root(IMAGENET_STD)

Thus, it turns out the image normalization is build into the model. When you perform this normalization yourself to the input image, the image will be normalized twice resulting in extremely small pixel values. The model will therefore have a hard time learning.

TLDR: Do not normalize the input image as it is build into the tf.keras.application model, input images should have values in the range 0-255.

Upvotes: 6

Related Questions