如何在 TensorFlow 2.0 中冻结 keras model？ （专门将一个保存的model格式冻结为.pb格式）

Question

Can any one please explain the procedure to freeze a keras model (saved model format) into.pb format in TensorFlow 2? Created a sample mobilenet keras model and saved it to the disk in saved model format

import tensorflow as tf
#Tensorflow version: 2.7.0
model = tf.keras.applications.mobilenet.MobileNet(
        include_top=True, 
        weights='imagenet', 
        input_tensor=None, 
        pooling=None,
        classes=1000
)
tf.keras.models.save_model( 
        model, 
        *path*, 
        overwrite=True, 
        include_optimizer=True, 
        save_format='pb', 
        signatures=None 
)

Then in another file, I need to load the model and freeze it to a.pb format

import tensorflow as tf
#Tensorflow version: 2.7.0
model = tf.keras.models.load_model( *path* )

############################################
# Freeze the model to a .pb format
############################################

Answer 1

With the advancement of tensorflow 2, freeze model has changed to saved model where instead of a single.pb file(graphdef), you now have saved model:

1. Weights
2. GraphDef (.pb)

import tensorflow as tf
pretrained_model = tf.keras.applications.MobileNet()
mobilenet_save_path = 'weights/mobilenet'

# Save to saved model
tf.saved_model.save(pretrained_model, mobilenet_save_path)

Note: Saved model format is faster and produce the exact same results

How to use saved model

import tensorflow as tf
model = tf.saved_model.load('weights/mobilenet/')

# Grab this function to run saved model
infer = model.signatures['serving_default'] 

image = 'something.jpg'
img = tf.io.decode_jpeg(tf.io.read_file(image))
img_pre = tf.cast(img, tf.float32) 
img_pre = (img_pre / 127.5) - 1
img_pre = tf.image.resize(img_pre, [224, 224])
img_pre = tf.expand_dims(img_pre, axis=0)
preds = infer(img_pre)['outputs']