root@14cab064210e:/xiesibo/tensorrt-demo-master# python trt_optimize.py
WARNING: Logging before flag parsing goes to stderr.
W0820 07:21:20.667007 140699697792768 deprecation_wrapper.py:119] From /usr/local/lib/python3.5/dist-packages/uff/converters/tensorflow/conversion_helpers.py:18: The name tf.GraphDef is deprecated. Please use tf.compat.v1.GraphDef instead.

W0820 07:21:20.667487 140699697792768 deprecation_wrapper.py:119] From /usr/local/lib/python3.5/dist-packages/uff/converters/tensorflow/conversion_helpers.py:231: The name tf.gfile.GFile is deprecated. Please use tf.io.gfile.GFile instead.

NOTE: UFF has been tested with TensorFlow 1.12.0. Other versions are not guaranteed to work
WARNING: The version of TensorFlow installed on this system is not guaranteed to work with UFF.
W0820 07:21:22.477570 140699697792768 deprecation_wrapper.py:119] From /usr/local/lib/python3.5/dist-packages/graphsurgeon/_utils.py:2: The name tf.NodeDef is deprecated. Please use tf.compat.v1.NodeDef instead.

UFF Version 0.6.3
=== Automatically deduced input nodes ===
[name: "net/input"
op: "Placeholder"
attr {
key: "dtype"
value {
type: DT_FLOAT
}
}
attr {
key: "shape"
value {
shape {
dim {
size: 1
}
dim {
size: 224
}
dim {
size: 224
}
dim {
size: 3
}
}
}
}
]

Using output node net/plate/BiasAdd
Using output node net/blet/BiasAdd
Converting to UFF graph
No. nodes: 92
[TensorRT] ERROR: UFFParser: Validator error: net/plate/BiasAdd: Unsupported operation Identity
***parser <tensorrt.tensorrt.UffParser object at 0x7ff6e677d0d8>
[TensorRT] ERROR: Network must have at least one output
Traceback (most recent call last):
File "trt_optimize.py", line 49, in
f.write(engine.serialize())
AttributeError: 'NoneType' object has no attribute 'serialize'

Хотел попробовать работу TensorRT, скачал датасет ILSVRC2012_img_train.tar по этому линку http://www.image-net.org/challenges/LSVRC/2012/nonpub-downloads , откуда можно скачать train.txt? Может есть смысл, добавить его в репозиторий

hi, I get a very low precision using tensorRT, but get 0.9580 using tf. here is my code:

def network(inputs, is_training=True, **kwargs):
#Lenet + BatchNorm
with tf.variable_scope('net'):
with slim.arg_scope([slim.conv2d, slim.fully_connected],
activation_fn=tf.nn.relu,
weights_initializer=tf.truncated_normal_initializer(0.0, 0.01),
weights_regularizer=slim.l2_regularizer(0.0005)):
with slim.arg_scope([slim.conv2d],
normalizer_fn=slim.batch_norm,
normalizer_params={'is_training': is_training}):
net = slim.conv2d(inputs, 32, [5, 5], scope='conv1')
net = slim.max_pool2d(net, [2, 2], scope='pool1')
net = slim.conv2d(net, 64, [5, 5], scope='conv2')
net = slim.max_pool2d(net, [2, 2], scope='pool2')
net = slim.flatten(net)
net = slim.fully_connected(net, 1024, scope='fc1')
net = slim.dropout(net, 0.5, scope='dropout', is_training=is_training)
net = slim.fully_connected(net, 10, activation_fn=None, scope='fc2')
return net

first i think you did a very great job,i am appreciate for it,however when i try to run the demo, i meet some problems like this: InvalidArgumentError (see above for traceback): LossTensor is inf or nan : Tensor had NaN values

[[Node: train_op/CheckNumerics = CheckNumerics[T=DT_FLOAT, _class=["loc:@train_op/control_dependency"], message="LossTensor is inf or nan", _device="/job:localhost/replica:0/task:0/device:GPU:0"](loss/sparse_softmax_cross_entropy_loss/value, ^net/conv5_2/BatchNorm/AssignMovingAvg_1, ^net/conv1_1/BatchNorm/AssignMovingAvg, ^net/conv1_1/BatchNorm/AssignMovingAvg_1, ^net/conv2_1/BatchNorm/AssignMovingAvg, ^net/conv2_1/BatchNorm/AssignMovingAvg_1, ^net/conv3_1/BatchNorm/AssignMovingAvg, ^net/conv3_1/BatchNorm/AssignMovingAvg_1, ^net/conv3_2/BatchNorm/AssignMovingAvg, ^net/conv3_2/BatchNorm/AssignMovingAvg_1, ^net/conv4_1/BatchNorm/AssignMovingAvg, ^net/conv4_1/BatchNorm/AssignMovingAvg_1, ^net/conv4_2/BatchNorm/AssignMovingAvg, ^net/conv4_2/BatchNorm/AssignMovingAvg_1, ^net/conv5_1/BatchNorm/AssignMovingAvg, ^net/conv5_1/BatchNorm/AssignMovingAvg_1, ^net/conv5_2/BatchNorm/AssignMovingAvg)]]
[[Node: train_op/control_dependency/_421 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_699_train_op/control_dependency", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
i know there might be some reason for it , i don't know what the reason is,so maybe you can help me ,thank you very much ,what's more,why the output node name is net/fc8/BiasAdd, i think the output node name might be conv5_2/BatchNorm/AssigMovingAvg ,plz tell me which one is the right answer and why,thanks