Skip to content

Instantly share code, notes, and snippets.

@zmjjmz

zmjjmz/export_error

Created August 14, 2019 21:17
Show Gist options
  • Star 0 You must be signed in to star a gist
  • Fork 0 You must be signed in to fork a gist
  • Save zmjjmz/1ec864404e4fed80e5ccf52c7589ea30 to your computer and use it in GitHub Desktop.
Save zmjjmz/1ec864404e4fed80e5ccf52c7589ea30 to your computer and use it in GitHub Desktop.
Download ZIP
Vespa Keras tf experiment export_saved_model issues
Raw
export_error
{'error-code': 'INVALID_APPLICATION_PACKAGE',
'message': 'Invalid application package: default.default: Error loading '
'model: Could not import TensorFlow model from directory '
"'/opt/vespa/var/db/vespa/config_server/serverdb/tenants/default/sessions/175/.preprocessed/models/plike_test/tf114_export': "
"_output_shapes attribute of 'init_1' does not exist"}
Raw
export_servonly_error
{'error-code': 'INVALID_APPLICATION_PACKAGE',
'message': 'Invalid application package: default.default: Error loading '
'model: Could not import TensorFlow model from directory '
"'/opt/vespa/var/db/vespa/config_server/serverdb/tenants/default/sessions/187/.preprocessed/models/plike_test/tf114_export_servonly': "
"NodeDef mentions attr 'config_proto' not in "
'Op<name=StatefulPartitionedCall; signature=args: -> output:; '
'attr=Tin:list(type),min=0; attr=Tout:list(type),min=0; '
'attr=f:func; is_stateful=true>; NodeDef: {{node '
'StatefulPartitionedCall_2}} = '
'StatefulPartitionedCall[Tin=[DT_STRING, DT_RESOURCE, DT_RESOURCE, '
'DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, '
'DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, '
'DT_RESOURCE], Tout=[DT_STRING], '
'_gradient_op_type="PartitionedCall-445", _output_shapes=[[]], '
'config_proto="\\n\\007\\n\\003GPU\\020\\000\\n\\007\\n\\003CPU\\020\\0012\\002J\\0008\\001", '
'f=__inference__traced_restore_444[], '
'_device="/job:localhost/replica:0/task:0/device:CPU:0"](_arg_saver_filename_0_0, '
'layer_1/kernel, layer_1/bias, output/kernel, output/bias, '
'RMSprop/iter, RMSprop/decay, RMSprop/learning_rate, '
'RMSprop/momentum, RMSprop/rho, RMSprop/layer_1/kernel/rms, '
'RMSprop/layer_1/bias/rms, RMSprop/output/kernel/rms, '
'RMSprop/output/bias/rms). (Check whether your '
'GraphDef-interpreting binary is up to date with your '
'GraphDef-generating binary.).\n'
'\t [[{{node StatefulPartitionedCall_2}} = '
'StatefulPartitionedCall[Tin=[DT_STRING, DT_RESOURCE, DT_RESOURCE, '
'DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, '
'DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, DT_RESOURCE, '
'DT_RESOURCE], Tout=[DT_STRING], '
'_gradient_op_type="PartitionedCall-445", _output_shapes=[[]], '
'config_proto="\\n\\007\\n\\003GPU\\020\\000\\n\\007\\n\\003CPU\\020\\0012\\002J\\0008\\001", '
'f=__inference__traced_restore_444[], '
'_device="/job:localhost/replica:0/task:0/device:CPU:0"](_arg_saver_filename_0_0, '
'layer_1/kernel, layer_1/bias, output/kernel, output/bias, '
'RMSprop/iter, RMSprop/decay, RMSprop/learning_rate, '
'RMSprop/momentum, RMSprop/rho, RMSprop/layer_1/kernel/rms, '
'RMSprop/layer_1/bias/rms, RMSprop/output/kernel/rms, '
'RMSprop/output/bias/rms)]]'}
Raw
export_servonly_signature
$ saved_model_cli show --dir tf114_export_servonly --all
MetaGraphDef with tag-set: 'serve' contains the following SignatureDefs:
signature_def['__saved_model_init_op']:
The given SavedModel SignatureDef contains the following input(s):
The given SavedModel SignatureDef contains the following output(s):
outputs['__saved_model_init_op'] tensor_info:
dtype: DT_INVALID
shape: unknown_rank
name: NoOp
Method name is:
signature_def['serving_default']:
The given SavedModel SignatureDef contains the following input(s):
inputs['input'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: serving_default_input:0
The given SavedModel SignatureDef contains the following output(s):
outputs['output'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: StatefulPartitionedCall:0
Method name is: tensorflow/serving/predict
Raw
export_signature
# note: set serving_only=False (which is the default) and it will export a lot of tag sets other than serve.
MetaGraphDef with tag-set: 'eval' contains the following SignatureDefs:
signature_def['__saved_model_init_op']:
The given SavedModel SignatureDef contains the following input(s):
The given SavedModel SignatureDef contains the following output(s):
outputs['__saved_model_init_op'] tensor_info:
dtype: DT_INVALID
shape: unknown_rank
name: init_1
Method name is:
signature_def['eval']:
The given SavedModel SignatureDef contains the following input(s):
inputs['input'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: input:0
inputs['output_target'] tensor_info:
dtype: DT_FLOAT
shape: (-1, -1)
name: output_target:0
The given SavedModel SignatureDef contains the following output(s):
outputs['loss'] tensor_info:
dtype: DT_FLOAT
shape: ()
name: loss/mul:0
outputs['predictions/output'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: output/BiasAdd:0
Method name is: tensorflow/supervised/eval
MetaGraphDef with tag-set: 'serve' contains the following SignatureDefs:
signature_def['__saved_model_init_op']:
The given SavedModel SignatureDef contains the following input(s):
The given SavedModel SignatureDef contains the following output(s):
outputs['__saved_model_init_op'] tensor_info:
dtype: DT_INVALID
shape: unknown_rank
name: init_1
Method name is:
signature_def['serving_default']:
The given SavedModel SignatureDef contains the following input(s):
inputs['input'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: input:0
The given SavedModel SignatureDef contains the following output(s):
outputs['output'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: output/BiasAdd:0
Method name is: tensorflow/serving/predict
MetaGraphDef with tag-set: 'train' contains the following SignatureDefs:
signature_def['__saved_model_init_op']:
The given SavedModel SignatureDef contains the following input(s):
The given SavedModel SignatureDef contains the following output(s):
outputs['__saved_model_init_op'] tensor_info:
dtype: DT_INVALID
shape: unknown_rank
name: init_1
Method name is:
signature_def['__saved_model_train_op']:
The given SavedModel SignatureDef contains the following input(s):
The given SavedModel SignatureDef contains the following output(s):
outputs['__saved_model_train_op'] tensor_info:
dtype: DT_INVALID
shape: unknown_rank
name: training_1/group_deps
Method name is:
signature_def['train']:
The given SavedModel SignatureDef contains the following input(s):
inputs['input'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: input:0
inputs['output_target'] tensor_info:
dtype: DT_FLOAT
shape: (-1, -1)
name: output_target:0
The given SavedModel SignatureDef contains the following output(s):
outputs['loss'] tensor_info:
dtype: DT_FLOAT
shape: ()
name: loss/mul:0
outputs['predictions/output'] tensor_info:
dtype: DT_FLOAT
shape: (-1, 1)
name: output/BiasAdd:0
Method name is: tensorflow/supervised/training
Raw
keras_export_saved_model.py
import tensorflow as tf
import numpy as np
input_l = tf.keras.Input(shape=(1,), name='input')
layer_1 = tf.keras.layers.Dense(1, activation='relu', name='layer_1')(input_l)
output_l = tf.keras.layers.Dense(1, activation='linear', name='output')(layer_1)
model = tf.keras.Model(inputs=[input_l], outputs=[output_l])
model.compile(loss='mean_absolute_error', optimizer='rmsprop')
x = np.array(np.arange(1, 100000))
y = np.array(np.arange(1, 100000))
model.fit(np.array(x).reshape(-1),np.array(y), epochs=2, shuffle=False, batch_size=100)
print("TF Version: {0}".format(tf.VERSION)) # should be 1.14
tf.keras.experimental.export_saved_model(
model,
'tf114_export_servonly',
serving_only=True
)
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment