TF2.2: Loading a Saved Model from tensorflow_hub failed with `AttributeError: '_UserObject' object has no attribute 'summary'`

Question

System info: Python: 3.6.9 Tensorflow: 2.2.0 CPU package from pip

The issue:

I got https://tfhub.dev/google/imagenet/resnet_v2_50/classification/4?tf-hub-format=compressed from tf-hub and then uncompressed in a new directory.

wget https://storage.googleapis.com/tfhub-modules/google/imagenet/resnet_v2_50/feature_vector/4.tar.gz
mkdir test_pb
mv 4.tar.gz test_pb
cd test_pb
tar -xvf 4.tar.gz
rm 4.tar.gz
cd ..
./test.py

https://storage.googleapis.com/tfhub-modules/google/imagenet/resnet_v2_50/feature_vector/4.tar.gz is the feature vector pre-trained model from https://tfhub.dev/google/imagenet/resnet_v2_50/feature_vector/4 in tf-hub.

test.py is the Python script, and this following is the standalone code:

#!/usr/bin/env python3
from __future__ import absolute_import, division, print_function, unicode_literals

import os
import tensorflow as tf

print(tf.__version__)

resnet50v2_save_path = os.path.join('.', "./test_pb/")

loaded1 = tf.keras.models.load_model(resnet50v2_save_path)
print("Load done")

print("Signatures: ", loaded1.signatures)

print("Type: ", type(loaded1))

print(loaded1.summary())

that give this output:

2.2.0
2020-06-12 20:29:07.677555: I tensorflow/core/platform/profile_utils/cpu_utils.cc:102] CPU Frequency: 1995455000 Hz
2020-06-12 20:29:07.678219: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x59eb130 initialized for platform Host (this does not guarantee that XLA will be used). Devices:
2020-06-12 20:29:07.678241: I tensorflow/compiler/xla/service/service.cc:176]   StreamExecutor device (0): Host, Default Version
Load done
Signatures:  _SignatureMap({})
Type:  ._UserObject'>
Traceback (most recent call last):
  File "./test.py", line 18, in 
    print(loaded1.summary())
AttributeError: '_UserObject' object has no attribute 'summary'

that is the mentioned error.

Why?

Thx

user_15460480 · Answer

The model is a plain SavedModel and cannot directly be loaded into a keras model since it lacks a keras_metadata.pb file. You have to load it like a normal SavedModel and wrap it in a keras model like so:

!wget https://storage.googleapis.com/tfhub-modules/google/imagenet/resnet_v2_50/feature_vector/4.tar.gz
!mkdir saved_model
!tar -xvf 4.tar.gz -C saved_model

import tensorflow as tf
import tensorflow_hub as hub

m = tf.keras.Sequential([hub.KerasLayer("saved_model", trainable=True),
    tf.keras.layers.Dense(10, activation='softmax')
])
m.build([None, 224, 224, 3])
m.summary()

# Model: "sequential"
# _________________________________________________________________
# Layer (type)                 Output Shape              Param #   
# =================================================================
# keras_layer (KerasLayer)     (None, 2048)              23564800  
# _________________________________________________________________
# dense (Dense)                (None, 10)                20490     
# =================================================================
# Total params: 23,585,290
# Trainable params: 23,539,850
# Non-trainable params: 45,440

TF2.2: Loading a Saved Model from tensorflow_hub failed with `AttributeError: '_UserObject' object has no attribute 'summary'`

1 Answers