46 lines
2.1 KiB
Python
46 lines
2.1 KiB
Python
# Copyright (C) 2018-2023 Intel Corporation
|
|
# SPDX-License-Identifier: Apache-2.0
|
|
|
|
import pytest
|
|
import tensorflow as tf
|
|
from common.tf_layer_test_class import CommonTFLayerTest
|
|
|
|
|
|
class TestIdentityN(CommonTFLayerTest):
|
|
def create_identityn_net(self, value_shape, size_splits_values, axis_value):
|
|
tf.compat.v1.reset_default_graph()
|
|
# Create the graph and model
|
|
with tf.compat.v1.Session() as sess:
|
|
axis = tf.constant(axis_value, dtype=tf.int32)
|
|
size_splits = tf.constant(size_splits_values, dtype=tf.int32)
|
|
value = tf.compat.v1.placeholder(tf.float32, value_shape, 'value')
|
|
num_split = len(size_splits_values)
|
|
# we have to do a trick using Split operation to avoid fusing output tensors into input tensors
|
|
# so that Parameter node will contain both input and output tensor names
|
|
# 97192: this is a limitation of layer test infrastructure
|
|
split = tf.raw_ops.SplitV(value=value, size_splits=size_splits, axis=axis, num_split=num_split)
|
|
split_outputs = []
|
|
for output_ind in range(num_split):
|
|
split_outputs.append(split[output_ind])
|
|
identity_n = tf.raw_ops.IdentityN(input=split_outputs)
|
|
for output_ind in range(num_split):
|
|
tf.identity(identity_n[output_ind], name="Identity_" + str(output_ind))
|
|
|
|
tf.compat.v1.global_variables_initializer()
|
|
tf_net = sess.graph_def
|
|
|
|
return tf_net, None
|
|
|
|
test_data_basic = [
|
|
dict(value_shape=[2, 3, 9], size_splits_values=[1, 2, 3, -1, 1], axis_value=2),
|
|
]
|
|
|
|
@pytest.mark.parametrize("params", test_data_basic)
|
|
@pytest.mark.precommit_tf_fe
|
|
@pytest.mark.nightly
|
|
def test_split_basic(self, params, ie_device, precision, ir_version, temp_dir,
|
|
use_new_frontend, use_old_api):
|
|
self._test(*self.create_identityn_net(**params),
|
|
ie_device, precision, ir_version, temp_dir=temp_dir,
|
|
use_new_frontend=use_new_frontend, use_old_api=use_old_api)
|