tensorflow / tensorrt Goto Github PK

TensorFlow/TensorRT integration

License: Apache License 2.0

Python 17.03% Shell 2.97% Jupyter Notebook 78.29% Starlark 0.04% C++ 1.49% CMake 0.19%

tensorrt's Introduction

Documentation for TensorRT in TensorFlow (TF-TRT)

TensorFlow-TensorRT (TF-TRT) is an integration of TensorFlow and TensorRT that leverages inference optimization on NVIDIA GPUs within the TensorFlow ecosystem. It provides a simple API that delivers substantial performance gains on NVIDIA GPUs with minimal effort. The documentation on how to accelerate inference in TensorFlow with TensorRT (TF-TRT) is here: https://docs.nvidia.com/deeplearning/dgx/tf-trt-user-guide/index.html

Check out this gentle introduction to TensorFlow TensorRT or watch this quick walkthrough example for more!

Examples for TensorRT in TensorFlow (TF-TRT)

This repository contains a number of different examples that show how to use TF-TRT. TF-TRT is a part of TensorFlow that optimizes TensorFlow graphs using TensorRT. We have used these examples to verify the accuracy and performance of TF-TRT. For more information see Verified Models.

Examples

Using TensorRT in TensorFlow (TF-TRT)

This module provides necessary bindings and introduces TRTEngineOp operator that wraps a subgraph in TensorRT. This module is under active development.

Installing TF-TRT

Currently Tensorflow nightly builds include TF-TRT by default, which means you don't need to install TF-TRT separately. You can pull the latest TF containers from docker hub or install the latest TF pip package to get access to the latest TF-TRT.

If you want to use TF-TRT on NVIDIA Jetson platform, you can find the download links for the relevant Tensorflow pip packages here: https://docs.nvidia.com/deeplearning/dgx/index.html#installing-frameworks-for-jetson

You can also use NVIDIA's Tensorflow container(tested and published monthly).

Installing TensorRT

In order to make use of TF-TRT, you will need a local installation of TensorRT from the NVIDIA Developer website. Installation instructions for compatibility with TensorFlow are provided on the TensorFlow GPU support guide.

Documentation

TF-TRT documentaion gives an overview of the supported functionalities, provides tutorials and verified models, explains best practices with troubleshooting guides.

Tests

TF-TRT includes both Python tests and C++ unit tests. Most of Python tests are located in the test directory and they can be executed uring bazel test or directly with the Python command. Most of the C++ unit tests are used to test the conversion functions that convert each TF op to a number of TensorRT layers.

Compilation

In order to compile the module, you need to have a local TensorRT installation (libnvinfer.so and respective include files). During the configuration step, TensorRT should be enabled and installation path should be set. If installed through package managers (deb,rpm), configure script should find the necessary components from the system automatically. If installed from tar packages, user has to set path to location where the library is installed during configuration.

bazel build --config=cuda --config=opt //tensorflow/tools/pip_package:build_pip_package
bazel-bin/tensorflow/tools/pip_package/build_pip_package /tmp/

License

Apache License 2.0

tensorrt's People

Contributors

Stargazers

Watchers

Forkers

github30 meldong pooyadavoodi trevor-m ritwik12 mdrozdowski1996 yuhonghong66 yuhonghong7035 anujonthemove davis-love-ai haohaibo fierval albertbj hxl1990 dreadlord1984 stevenlol haikuoyao tuanho27 jweiss63 ttyio pcicotti aaroey dou8856 monklof lyimage mpjlu blairhan firefoxhtjc david-imagr otstrel xiejingcai kristenxy gavinljj cscn89 oscarriddle nikithamuddireddy ashwin artchula baking-soda muralir-nv tjakob egerszegilaszlo xkx0430 kewlcoder akrusher vinhngx ersks esmeyi huaifeng1993 pu2wof butterflydisappeared cathy-kim koc4466677 pidem mbrukman ddavila-kitware eycab pabgonza phillip-kravtsov jetson-tx2 platoaria nluehr ca-kamiya-joji aminehy jyotendra argonarnav lanchongyizu stereoboy float123 walterchen08 dominicshanshan soldierofhell tonycao gypsybud gjhanchem ohlr npanpaliya hubbucket-team adrianosantospb phymucs leo-xukang cristinamelicio s-koksharova shuaihugao michaelliu-tj salluru007 gaowq2017 jamesguo zuojianhao prasanthpul animohan dlt1995 wwy510553871 baoxuhong tfresources cloudrivers helenyou chuhai-lin jarain78 suhyung-code42

tensorrt's Issues

image_classification test fails

Trying to run the examples in tensorrt fails with the following error

python image_classification.py

Traceback (most recent call last):
File "image_classification.py", line 30, in
import official.resnet.imagenet_main
File "/root/antojohn/tensorrt/tftrt/examples/third_party/models/official/resnet/imagenet_main.py", line 31, in
from official.resnet import resnet_run_loop
File "/root/antojohn/tensorrt/tftrt/examples/third_party/models/official/resnet/resnet_run_loop.py", line 41, in
from official.utils.misc import distribution_utils
File "/root/antojohn/tensorrt/tftrt/examples/third_party/models/official/utils/misc/distribution_utils.py", line 28, in
None: tf.distribute.experimental.CollectiveCommunication.AUTO,
AttributeError: 'module' object has no attribute 'experimental'

Any thoughts/suggestions ?

You must feed a value for placeholder tensor 'import/input_tensor' with dtype string and shape

first, I download resnet_v2_fp32_savedmodel_NHWC_jpg saved_model from blog

then, I run a example provided in nvidia documents about trt-tf here :

# Import TensorFlow and TensorRT
import tensorflow as tf
import tensorflow.contrib.tensorrt as trt
# Inference with TF-TRT `SavedModel` workflow:
graph = tf.Graph()
with graph.as_default():
    with tf.Session() as sess:
        # Create a TensorRT inference graph from a SavedModel:
        trt_graph = trt.create_inference_graph(
            input_graph_def=None,
            outputs=None,
            input_saved_model_dir="/tmp/resnet/1538687457",
            input_saved_model_tags=["serve"]) 
        # Import the TensorRT graph into a new graph and run:
        output_node = tf.import_graph_def(
            trt_graph,
            return_elements=["softmax_tensor"])
        sess.run(output_node)

but it shows error:

2019-03-19 17:21:53.332397: I tensorflow/contrib/tensorrt/convert/convert_graph.cc:913] Number of TensorRT candidate segments: 1
2019-03-19 17:21:54.040844: W tensorflow/contrib/tensorrt/convert/convert_nodes.cc:3710] Validation failed for TensorRTInputPH_0 and input slot 0: Input tensor with shape [?,?,?,3] has an unknown non-batch dimension at dim 1
2019-03-19 17:21:54.040922: W tensorflow/contrib/tensorrt/convert/convert_graph.cc:1021] TensorRT node TRTEngineOp_0 added for segment 0 consisting of 445 nodes failed: Invalid argument: Validation failed for TensorRTInputPH_0 and input slot 0: Input tensor with shape [?,?,?,3] has an unknown non-batch dimension at dim 1. Fallback to TF...
2019-03-19 17:21:54.289076: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:581] Optimization results for grappler item: tf_graph
2019-03-19 17:21:54.289142: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   constant folding: Graph size after: 550 nodes (-256), 613 edges (-258), time = 820.515ms.
2019-03-19 17:21:54.289160: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   layout: Graph size after: 557 nodes (7), 615 edges (2), time = 142.796ms.
2019-03-19 17:21:54.289171: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   constant folding: Graph size after: 552 nodes (-5), 615 edges (0), time = 535.989ms.
2019-03-19 17:21:54.289184: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   TensorRTOptimizer: Graph size after: 552 nodes (0), 615 edges (0), time = 1296.37695ms.
2019-03-19 17:22:03.134267: I tensorflow/stream_executor/dso_loader.cc:152] successfully opened CUDA library libcublas.so.10.0 locally
Traceback (most recent call last):
  File "trt-tf.py", line 18, in <module>
    sess.run(output_node)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 929, in run
    run_metadata_ptr)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1152, in _run
    feed_dict_tensor, options, run_metadata)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1328, in _do_run
    run_metadata)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/client/session.py", line 1348, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: You must feed a value for placeholder tensor 'import/input_tensor' with dtype string and shape [?]
	 [[node import/input_tensor (defined at trt-tf.py:17) ]]
	 [[node import/map/Shape (defined at trt-tf.py:17) ]]

Caused by op u'import/input_tensor', defined at:
  File "trt-tf.py", line 17, in <module>
    return_elements=["softmax_tensor"])
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/util/deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/framework/importer.py", line 442, in import_graph_def
    _ProcessNewOps(graph)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/framework/importer.py", line 235, in _ProcessNewOps
    for new_op in graph._add_new_tf_operations(compute_devices=False):  # pylint: disable=protected-access
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 3433, in _add_new_tf_operations
    for c_op in c_api_util.new_tf_operations(self)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 3325, in _create_op_from_tf_operation
    ret = Operation(c_op, self)
  File "/xxx/.virtualenvs/tf-tmp/lib/python2.7/site-packages/tensorflow/python/framework/ops.py", line 1801, in __init__
    self._traceback = tf_stack.extract_stack()

InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'import/input_tensor' with dtype string and shape [?]
	 [[node import/input_tensor (defined at trt-tf.py:17) ]]
	 [[node import/map/Shape (defined at trt-tf.py:17) ]]

how to fix it? anyone can give some advises?

my envs:
tensorflow-gpu 1.13.1

Does `tensorrt` module is needed for c binding for inference ?

We compiled tensorflow with and without tensorrt for jetson (arm64). But now when I'm thinking on it, does tensorrt is actually needed for inference stage?

Segmentation fault when optimize model.

Currently I'm trying with ssd_mobilenet_v2_coco with an NVIDIA 1060GTX.

I have tensorflow-gpu v1.13., CUDA10. TensorRT 5.
I've downloaded the model with

config_path, checkpoint_path = download_model('ssd_mobilenet_v2_coco', output_dir='models')

I'm trying to optimize the model. with


frozen_graph = optimize_model(
    config_path=config_path, 
    checkpoint_path=checkpoint_path,
    use_trt=True,
    precision_mode='FP16'
)

But always provokes a segmentation fault.. this is the log console.

2019-03-20 11:26:09.152490: I tensorflow/core/platform/cpu_feature_guard.cc:141] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA 2019-03-20 11:26:09.235499: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero 2019-03-20 11:26:09.235934: I tensorflow/compiler/xla/service/service.cc:150] XLA service 0x45ea9f0 executing computations on platform CUDA. Devices: 2019-03-20 11:26:09.235950: I tensorflow/compiler/xla/service/service.cc:158] StreamExecutor device (0): GeForce GTX 1060 6GB, Compute Capability 6.1 2019-03-20 11:26:09.257165: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 2808000000 Hz 2019-03-20 11:26:09.257710: I tensorflow/compiler/xla/service/service.cc:150] XLA service 0x3d6e750 executing computations on platform Host. Devices: 2019-03-20 11:26:09.257725: I tensorflow/compiler/xla/service/service.cc:158] StreamExecutor device (0): <undefined>, <undefined> 2019-03-20 11:26:09.257948: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1433] Found device 0 with properties: name: GeForce GTX 1060 6GB major: 6 minor: 1 memoryClockRate(GHz): 1.7845 pciBusID: 0000:01:00.0 totalMemory: 5.93GiB freeMemory: 5.56GiB 2019-03-20 11:26:09.257964: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0 2019-03-20 11:26:09.334021: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix: 2019-03-20 11:26:09.334056: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0 2019-03-20 11:26:09.334062: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0: N 2019-03-20 11:26:09.334197: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5369 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1) WARNING:tensorflow:From /home/idiaz/.local/lib/python3.6/site-packages/object_detection-0.1-py3.6.egg/object_detection/exporter.py:327: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version. Instructions for updating: Use tf.cast instead. WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/op_def_library.py:263: colocate_with (from tensorflow.python.framework.ops) is deprecated and will be removed in a future version. Instructions for updating: Colocations handled automatically by placer. INFO:tensorflow:depth of additional conv before box predictor: 0 INFO:tensorflow:depth of additional conv before box predictor: 0 INFO:tensorflow:depth of additional conv before box predictor: 0 INFO:tensorflow:depth of additional conv before box predictor: 0 INFO:tensorflow:depth of additional conv before box predictor: 0 INFO:tensorflow:depth of additional conv before box predictor: 0 WARNING:tensorflow:From /home/idiaz/.local/lib/python3.6/site-packages/object_detection-0.1-py3.6.egg/object_detection/exporter.py:356: get_or_create_global_step (from tensorflow.contrib.framework.python.ops.variables) is deprecated and will be removed in a future version. Instructions for updating: Please switch to tf.train.get_or_create_global_step 2019-03-20 11:26:15.423701: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0 2019-03-20 11:26:15.423744: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix: 2019-03-20 11:26:15.423750: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0 2019-03-20 11:26:15.423753: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0: N 2019-03-20 11:26:15.423886: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5369 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1) WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/training/saver.py:1266: checkpoint_exists (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version. Instructions for updating: Use standard file APIs to check for files with this prefix. INFO:tensorflow:Restoring parameters from models/ssd_mobilenet_v2_coco_2018_03_29/model.ckpt 2019-03-20 11:26:18.529755: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0 2019-03-20 11:26:18.529812: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix: 2019-03-20 11:26:18.529820: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0 2019-03-20 11:26:18.529825: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0: N 2019-03-20 11:26:18.529932: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5369 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1) INFO:tensorflow:Restoring parameters from models/ssd_mobilenet_v2_coco_2018_03_29/model.ckpt WARNING:tensorflow:From /home/idiaz/.local/lib/python3.6/site-packages/object_detection-0.1-py3.6.egg/object_detection/exporter.py:96: convert_variables_to_constants (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version. Instructions for updating: Use tf.compat.v1.graph_util.convert_variables_to_constants WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/framework/graph_util_impl.py:245: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version. Instructions for updating: Use tf.compat.v1.graph_util.extract_sub_graph INFO:tensorflow:Froze 344 variables. INFO:tensorflow:Converted 344 variables to const ops. 2019-03-20 11:26:19.787213: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0 2019-03-20 11:26:19.787255: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix: 2019-03-20 11:26:19.787261: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0 2019-03-20 11:26:19.787264: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0: N 2019-03-20 11:26:19.787369: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5369 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1) WARNING:tensorflow:From /home/idiaz/.local/lib/python3.6/site-packages/object_detection-0.1-py3.6.egg/object_detection/exporter.py:288: build_tensor_info (from tensorflow.python.saved_model.utils_impl) is deprecated and will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info. WARNING:tensorflow:From /home/idiaz/.local/lib/python3.6/site-packages/object_detection-0.1-py3.6.egg/object_detection/exporter.py:288: build_tensor_info (from tensorflow.python.saved_model.utils_impl) is deprecated and will be removed in a future version. Instructions for updating: This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info. INFO:tensorflow:No assets to save. INFO:tensorflow:No assets to save. INFO:tensorflow:No assets to write. INFO:tensorflow:No assets to write. INFO:tensorflow:SavedModel written to: .optimize_model_tmp_dir/saved_model/saved_model.pb INFO:tensorflow:SavedModel written to: .optimize_model_tmp_dir/saved_model/saved_model.pb INFO:tensorflow:Writing pipeline config file to .optimize_model_tmp_dir/pipeline.config INFO:tensorflow:Writing pipeline config file to .optimize_model_tmp_dir/pipeline.config 2019-03-20 11:26:21.916570: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0 2019-03-20 11:26:21.916607: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix: 2019-03-20 11:26:21.916613: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0 2019-03-20 11:26:21.916617: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0: N 2019-03-20 11:26:21.916717: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5369 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1) INFO:tensorflow:Running against TensorRT version 5.0.2 INFO:tensorflow:Running against TensorRT version 5.0.2 2019-03-20 11:26:23.734739: I tensorflow/core/grappler/devices.cc:51] Number of eligible GPUs (core count >= 8): 1 2019-03-20 11:26:23.735758: I tensorflow/core/grappler/clusters/single_machine.cc:359] Starting new session 2019-03-20 11:26:23.738573: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0 2019-03-20 11:26:23.738598: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix: 2019-03-20 11:26:23.738603: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990] 0 2019-03-20 11:26:23.738607: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0: N 2019-03-20 11:26:23.738711: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 5369 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1060 6GB, pci bus id: 0000:01:00.0, compute capability: 6.1) 2019-03-20 11:26:24.093794: W tensorflow/core/framework/allocator.cc:124] Allocation of 25159680 exceeds 10% of system memory. 2019-03-20 11:26:24.117771: W tensorflow/core/framework/allocator.cc:124] Allocation of 25159680 exceeds 10% of system memory. 2019-03-20 11:26:24.219573: W tensorflow/core/framework/allocator.cc:124] Allocation of 25159680 exceeds 10% of system memory. 2019-03-20 11:26:24.219672: W tensorflow/core/framework/allocator.cc:124] Allocation of 25159680 exceeds 10% of system memory. 2019-03-20 11:26:24.435736: W tensorflow/core/framework/allocator.cc:124] Allocation of 25159680 exceeds 10% of system memory. 2019-03-20 11:26:25.790160: I tensorflow/contrib/tensorrt/segment/segment.cc:443] There are 2317 ops of 33 different types in the graph that are not converted to TensorRT: Fill, Switch, TopKV2, ConcatV2, Identity, Squeeze, Const, Unpack, ResizeBilinear, Reshape, Mul, Slice, Merge, Split, NonMaxSuppressionV3, GatherV2, Range, Conv2D, Cast, Greater, Minimum, Sub, StridedSlice, NoOp, ZerosLike, Pack, Transpose, ExpandDims, Where, Exp, Placeholder, Add, Shape, (For more information see https://docs.nvidia.com/deeplearning/dgx/integrate-tf-trt/index.html#support-ops). 2019-03-20 11:26:26.231074: I tensorflow/contrib/tensorrt/convert/convert_graph.cc:913] Number of TensorRT candidate segments: 187 2019-03-20 11:26:35.074128: I tensorflow/contrib/tensorrt/convert/convert_graph.cc:1015] TensorRT node TRTEngineOp_0 added for segment 0 consisting of 224 nodes succeeded. 2019-03-20 11:26:35.074828: W tensorflow/contrib/tensorrt/convert/convert_graph.cc:1021] TensorRT node BoxPredictor_1/ClassPredictor/TRTEngineOp_1 added for segment 1 consisting of 2 nodes failed: Internal: Segment has no inputs (possible constfold failure). Fallback to TF... Segmentation fault (core dumped)

Broken link in README.md

Link for Image Classification appears to be broken.

inference with tf-trt

During convert frozen model to tf-trt optimized model, i have some question:

I convert model use tensorrt function create_inference_graph, after that i print graph's node name and node op, and found there is no op named 'TRTEngineOp', is this means the optimize failed? code like this:

with tf.Session() as sess:
    # read a frozen model:
    with tf.gfile.GFile(sys.argv[1], 'rb') as f:
        graph_def = tf.GraphDef()
        graph_def.ParseFromString(f.read())
    trt_graph = trt.create_inference_graph(
        input_graph_def = graph_def,
        outputs = outstr,
        max_batch_size = 1,
        max_workspace_size_bytes = (1<<20),
        precision_mode = "FP32")
    with open('trt.op', 'w') as fw:
        for n in trt_graph.node:
            fw.write(n.op+'\t'+n.name+'\n')
    constant_graph = graph_util.convert_variables_to_constants(sess, trt_graph, outstr)
    with tf.gfile.GFile(sys.argv[2], "wb") as f:
        f.write(constant_graph.SerializeToString())

Can i save the tf-trt optimized model(like .pb) and run the optimized model with cpp for online inference?
Is there a way or example for using tf-trt with cpp for model inference, i have already compile tensorflow with tensorrt from source, but i can't find any demo or api for cpp inference using tf-trt.
Thanks.

Problems in benchmark_model when use_synthetic=True.

There seem to be a couple of errors in benchmark_model() that appear when use_synthetic=True.

I get TypeError: 'module' object is not callable when calling benchmark_model with use_synthetic=True and where image_shape is given, due to an error at

tensorrt/tftrt/examples/object_detection/object_detection.py

Line 579 in 2b07cdb

batch_images = np.random(256, size=(batch_size, image_shape[0], image_shape[1], 3))

The previous line correctly uses np.random.randint but this line uses np.random.
Additionally, the image_ids variable is not set properly leading to an UnboundLocalError at

tensorrt/tftrt/examples/object_detection/object_detection.py

Line 611 in 2b07cdb

(len(image_ids) + batch_size - 1) / batch_size,

The following changes seem to fix these problems.

diff --git a/tftrt/examples/object_detection/object_detection.py b/tftrt/examples/object_detection/object_detection.py
index 2736199..50179ec 100644
--- a/tftrt/examples/object_detection/object_detection.py
+++ b/tftrt/examples/object_detection/object_detection.py
@@ -531,7 +531,9 @@ def benchmark_model(frozen_graph,
         raise RuntimeError(
             'Fixed image shape must be provided for batch size > 1')

-    if not use_synthetic:
+    if use_synthetic:
+        image_ids = list(range(num_images))
+    else:
         coco = COCO(annotation_file=annotation_path)


@@ -576,7 +578,7 @@ def benchmark_model(frozen_graph,
                     if image_shape is None:
                         batch_images = np.random.randint(256, size=(batch_size, 600, 600, 3))
                     else:
-                        batch_images = np.random(256, size=(batch_size, image_shape[0], image_shape[1], 3))
+                        batch_images = np.random.randint(256, size=(batch_size, image_shape[0], image_shape[1], 3))
                 else:
                     batch_image_ids = image_ids[image_idx:image_idx + batch_size]
                     batch_images = []

I could also fork and create a pull request if that will be easier.

TRT Conversion from TF saves dictionary file in graph making the protobuf too large

Hi,

To produce take any model and add a 1GB or some large dictionary in the assets folder for a saved_model format. Run the TF-TRT conversion which will produce a protobuf file of size > 1GB or sometimes 2GB. If we then try to serve the model through TF-Serving or NVIDIA TRT Server then you will get this error:

A protocol message was rejected because it was too big (more than 1073741824 bytes). To increase the limit (or to disable these warnings), see CodedInputStream::SetTotalBytesLimit() in google/protobuf/io/coded_stream.h.

Can we specify the TF-TRT converter to not save the dictionary in the graph and load it when the model is loaded?

Thanks!

Unable to run bigger batch sizes

Hi,

I am unable to run batchsizes larger than 1000. Has anyone else faced the same problem?

TRT inference with graph trained in a "FIFO queue" style

Hi,

Bear with me if I ask this question in the wrong place.

I am trying to optimize a tf model with TRT. I have encountered some trouble:

1.Original model is trained in a FIFO queue style. So basically [image, label, existence] are pushed into a queue and this bunch of data is dequeued before forward propagation.

2.I loaded the trained checkpoint and freeze it to a frozen graph. Frozen graph is optimized using trt_create_inference_graph().

I want to use get_tenor_by_name to get the input and 2 outputs and use sess.run() to make a prediction. And this fails. I have printed out the nodes and found the pre-processing nodes are removed and only a queue node remains like this:

name: "prefetch_queue/fifo_queue"
op: "FIFOQueueV2"
attr {
key: "capacity"
value {
i: 4
}
}
attr {
key: "component_types"
value {
list {
type: DT_FLOAT
type: DT_INT32
type: DT_FLOAT
}
}
}
attr {
key: "container"
value {
s: ""
}
}
attr {
key: "shapes"
value {
list {
shape {
dim {
size: 4
}
dim {
size: 256
}
dim {
size: 512
}
dim {
size: 3
}
}
shape {
dim {
size: 4
}
dim {
size: 256
}
dim {
size: 512
}
}
shape {
dim {
size: 4
}
dim {
size: 4
}
}
}
}
}
attr {
key: "shared_name"
value {
s: ""
}
}

I wanted to replace the queue with a placeholder (?,256,512,3) and run the model based on the optimized graph but there is no luck so far.
(Failed try:
#freezed_input_tensor has name INPUT+':0'
output1 = get_tensor_by_name(output1+':0')
output2 = get_tensor_by_name(output2+':0')
my_input_tensor = tf.placeholder(dtype=tf.float32, shape=[1,256,512,3], name='my_input')
tf.import_graph_def(trt_graph, input_map = {INPUT + ':0':my_input_tensor},name='')
sess.run(output1,output2, feed_dict:{my_input_tensor : my_image}) -->>fails at this step)

As the code provided only shows how inference can be optimised, could you please show how the model is trained as well?

Also, has anybody encountered the situation where I want to feed a placeholder to a graph trained with a FIFO queue?

Best,

0 trt_only Nodes for ssd_mobilenet_v1_coco

I followed the documentation steps here.

(another issue is that the current documentation provides a checkpoint_path to optimize_model, but optimize_model only takes a frozen graph. I fixed this by loading the frozen graph in and passing it to optimize_model.

This is the output from the optimize_model call (following the example):

graph_size(MB)(native_tf): 27.8
graph_size(MB)(trt): 27.8
num_nodes(native_tf): 6063
num_nodes(tftrt_total): 6047
num_nodes(trt_only): 0
time(s) (trt_conversion): 4.6702

Is this expected? Why are there zero trt_only nodes?

how to export tensorrt int8 model with savedmodel format after calibration for tf-serving？

as the latest doc shows, we can call create_inference_graph with parameter input_saved_model_dir and output_saved_model_dir to export a savedModel model.
If use int8 qutization，it seems that after create_inference_graph，we still need call calib_graph_to_infer_graph to export the final model and the result is a frozen model.
So...

Is it possible to export a savedModel int8 model after calibration for tf-serving?
If it is possible, is there any demo could be offered？
If it is impossible，when will it may be possible？

Segmentation Fault during Inference

During inference I encounter the following:

boxes, classes, scores, num_detections = tf_sess.run([tf_boxes, tf_classes, tf_scores, tf_num_detections], feed_dict={tf_input: image_resized[None,...]})

2019-08-13 14:20:15.309738: W tensorflow/compiler/jit/mark_for_compilation_pass.cc:1412] (One-time warning): Not using XLA:CPU for cluster because envvar TF_XLA_FLAGS=--tf_xla_cpu_global_jit was not set. If you want XLA:CPU, either set that envvar, or use experimental_jit_scope to enable XLA:CPU. To confirm that XLA is active, pass --vmodule=xla_compilation_cache=1 (as a proper command-line flag, not via TF_XLA_FLAGS) or set the envvar XLA_FLAGS=--xla_hlo_profile.
2019-08-13 14:20:15.384148: I tensorflow/compiler/tf2tensorrt/kernels/trt_engine_op.cc:632] Building a new TensorRT engine for TRTEngineOp_0 input shapes: [[1,500,500,3]]
2019-08-13 14:20:15.401345: E tensorflow/compiler/tf2tensorrt/utils/trt_logger.cc:41] DefaultLogger (Unnamed Layer* 3) [Convolution]: at least three non-batch dimensions are required for input
2019-08-13 14:20:15.401649: E tensorflow/compiler/tf2tensorrt/utils/trt_logger.cc:41] DefaultLogger (Unnamed Layer* 9) [Convolution]: at least three non-batch dimensions are required for input
Segmentation fault (core dumped)

Any help will be appreciated

Thanks

how to deal with the whileloop or tensorarray in TF model

hi,
SSD postproessor combined with while loop and tensorarray, how tensorRT deal with control flow op?
enter merge switch etc.
how to allocate memory for tensorarray object.

These op confused me for a long time. Truthly thanks!

run_calibration argument not used in run function

The run function is in tensorrt/tftrt/examples/image-classification/image_classification.py

create_inference_graph() causing "placeholder tensor" issue

Hi,

I'm running into the same issue as #44 where using create_inference_graph() returns error "You must feed a value for placeholder tensor 'import/input_tensor' with dtype string and shape".

I've had this happen with my own frozen graph, as well as the graph mentioned in #44 .

That issue has been closed, but there is no resolution. What needs to be done?

I'm running Tensorflow-GPU 1.14.0 on ubuntu 16 with Cuda 10.0.

Thanks in advance,

My error message:

tensorflow.python.framework.errors_impl.InvalidArgumentError: You must feed a value for placeholder tensor 'import/input_tensor' with dtype string and shape [?]
[[node import/input_tensor (defined at convert.py:19) ]]

No tensorrt model in win10 python3.5 and python3.6

Exause Me
I use tensorflow in windows10, python3.5
when I do
"import tensorflow.contrib.tensorrt as trt"
error appears:
ImportError: No module named 'tensorflow.contrib.tensorrt'

How can I get the tensorrt model in tensorflow for win10 ?

the result of sample end_to_end_tensorflow_mnist

TypeError: optimize_model() got an unexpected keyword argument 'config_path'

Hi,

TypeError: optimize_model() got an unexpected keyword argument 'config_path'

I got above error when following the tutorial below

frozen_graph = optimize_model(
config_path=config_path,
...)

The command interface refers to r1.13 rather than the master branch
def optimize_model(frozen_graph,
use_trt=True,
...)

r1.13
frozen_graph = optimize_model(
config_path=config_path,
checkpoint_path=checkpoint_path,
...)

InvalidArgumentError: Expected image (JPEG, PNG, or GIF)

I run Image classification example but it taks errors:

the ImageNet validation data like this，generate with the build_imagenet_data/build_imagenet_data.py

TensorRT with multiple models crash down.

The error message is:
ValueError: Cannot add function 'my_trt_op_1_native_segment' because a different function with the same name already exists.

The code (I put it here: https://github.com/Mengjintao/TF_TensorRT_test) will generate the above error message. But occasionally it may works fine.

I have no idea why this happens when I use multiple models with TF over TensorRT. Our code TF over CUDNN works fine.

Jintao Meng
2019.5.15

[Object Detection] ‘faster_rcnn_resnet50_coco’ model is errored during inference, after trt optimization: 'Incompatible shapes: [17328] vs. [16150]'

GPU: 1080Ti
docker: nvcr.io/nvidia/tensorflow 19.03-py3 e8ec6af2fb6f

My test.py:


from tftrt.examples.object_detection import download_model
                                                                                       
config_path, checkpoint_path = download_model('faster_rcnn_resnet50_coco', output_dir='models')
from tftrt.examples.object_detection import optimize_model

frozen_graph = optimize_model(
    config_path=config_path,
    checkpoint_path=checkpoint_path,
    use_trt=True,
    precision_mode='FP32',
    calib_images_dir='dataset/calib_dir'
)

from tftrt.examples.object_detection import download_dataset

images_dir, annotation_path = download_dataset('val2014', output_dir='dataset')

from tftrt.examples.object_detection import benchmark_model

statistics = benchmark_model(
    frozen_graph=frozen_graph,
    images_dir=images_dir,
    annotation_path=annotation_path
)

print('avg_latency_ms: ',statistics['avg_latency_ms'])
print('avg_throughput_fps: ',statistics['avg_throughput_fps'])
print('map: ',statistics['map'])

from tftrt.examples.object_detection import test

LOG：


2019-04-29 09:34:10.987362: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:7 with 10377 MB memory) -> physical GPU (device: 7, name: GeForce GTX 1080 Ti, pci bus id: 0000:8a:00.0, compute capability: 6.1)
  0%|                                                                                                                                                           | 0/4096 [00:00<?, ?it/s]2019-04-29 09:34:15.090991: W ./tensorflow/core/grappler/optimizers/graph_optimizer_stage.h:241] Failed to run optimizer ArithmeticOptimizer, stage RemoveStackStridedSliceSameAxis node Preprocessor/ResizeToRange/strided_slice_3. Error: Pack node (Preprocessor/ResizeToRange/stack_2) axis attribute is out of bounds: 0
2019-04-29 09:34:15.785398: W ./tensorflow/core/grappler/optimizers/graph_optimizer_stage.h:241] Failed to run optimizer ArithmeticOptimizer, stage RemoveStackStridedSliceSameAxis node Preprocessor/ResizeToRange/strided_slice_3. Error: Pack node (Preprocessor/ResizeToRange/stack_2) axis attribute is out of bounds: 0
2019-04-29 09:34:16.334620: I tensorflow/stream_executor/dso_loader.cc:153] successfully opened CUDA library libcublas.so.10 locally
2019-04-29 09:34:16.598452: I tensorflow/contrib/tensorrt/kernels/trt_engine_op.cc:496] Building a new TensorRT engine for TRTEngineOp_0 with batch size 1
2019-04-29 09:34:38.195497: I tensorflow/contrib/tensorrt/kernels/trt_engine_op.cc:496] Building a new TensorRT engine for TRTEngineOp_1 with batch size 100
  2%|██▍                                                                                                                                               | 67/4096 [00:32<05:14, 12.79it/s]Traceback (most recent call last):
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 1334, in _do_call
    return fn(*args)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 1319, in _run_fn
    options, feed_dict, fetch_list, target_list, run_metadata)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 1407, in _call_tf_sessionrun
    run_metadata)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [17328] vs. [16150]
	 [[{{node Decode/mul_3}}]]
	 [[{{node SecondStagePostprocessor/Reshape_4}}]]

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "test.py", line 24, in <module>
    annotation_path=annotation_path
  File "/opt/tensorflow/nvidia-examples/tensorrt/tftrt/examples/object_detection/object_detection.py", line 568, in benchmark_model
    feed_dict={tf_input: batch_images})
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 929, in run
    run_metadata_ptr)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 1152, in _run
    feed_dict_tensor, options, run_metadata)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 1328, in _do_run
    run_metadata)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/client/session.py", line 1348, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: Incompatible shapes: [17328] vs. [16150]
	 [[node Decode/mul_3 (defined at /opt/tensorflow/nvidia-examples/tensorrt/tftrt/examples/object_detection/object_detection.py:535) ]]
	 [[node SecondStagePostprocessor/Reshape_4 (defined at /opt/tensorflow/nvidia-examples/tensorrt/tftrt/examples/object_detection/object_detection.py:535) ]]

Caused by op 'Decode/mul_3', defined at:
  File "test.py", line 24, in <module>
    annotation_path=annotation_path
  File "/opt/tensorflow/nvidia-examples/tensorrt/tftrt/examples/object_detection/object_detection.py", line 535, in benchmark_model
    tf.import_graph_def(frozen_graph, name='')
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/util/deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/importer.py", line 442, in import_graph_def
    _ProcessNewOps(graph)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/importer.py", line 235, in _ProcessNewOps
    for new_op in graph._add_new_tf_operations(compute_devices=False):  # pylint: disable=protected-access
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/ops.py", line 3433, in _add_new_tf_operations
    for c_op in c_api_util.new_tf_operations(self)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/ops.py", line 3433, in <listcomp>
    for c_op in c_api_util.new_tf_operations(self)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/ops.py", line 3325, in _create_op_from_tf_operation
    ret = Operation(c_op, self)
  File "/usr/local/lib/python3.5/dist-packages/tensorflow/python/framework/ops.py", line 1801, in __init__
    self._traceback = tf_stack.extract_stack()

InvalidArgumentError (see above for traceback): Incompatible shapes: [17328] vs. [16150]
	 [[node Decode/mul_3 (defined at /opt/tensorflow/nvidia-examples/tensorrt/tftrt/examples/object_detection/object_detection.py:535) ]]
	 [[node SecondStagePostprocessor/Reshape_4 (defined at /opt/tensorflow/nvidia-examples/tensorrt/tftrt/examples/object_detection/object_detection.py:535) ]]

I noticed that this error did not happen at the very first test image, but after running a few of images (67 images finished). Is it a bug?

Does not work on FPN model?

Hi,

I have a self-implemented object detection model here:

Basically it is a Resnet101 + FPN + Faster RCNN model.

I used this code with these commands to convert the frozen graph and it reports no error:

graph_size(MB)(native_tf): 245.5
graph_size(MB)(trt): 245.0
num_nodes(native_tf): 5125
num_nodes(tftrt_total): 2618
num_nodes(trt_only): 47
time(s) (trt_conversion): 19.7020

Then during inference, I got this error:

InternalError (see above for traceback): Native FunctionDef fpn/upsample_lat5/Tensordot/TRTEngineOp_39_native_segment can't be found in function library
[[node model_0/fpn/upsample_lat5/Tensordot/TRTEngineOp_39 (defined at /home/junweil/object_detection/script/tf_mrcnn/models.py:147) ]]
[[{{node model_0/fastrcnn_predictions/map/while/body/_1/GatherV2_1}}]]

Could you give me some pointers? Thanks!

map does not match the numbers in models project

download the project and add project root to $PYTHONPATH.
I add a line to bypass optimization step:

diff --git a/tftrt/examples/object_detection/object_detection.py b/tftrt/examples/object_detection/object_detection.py
index 0d3528c..05b6cd7 100644
--- a/tftrt/examples/object_detection/object_detection.py
+++ b/tftrt/examples/object_detection/object_detection.py
@@ -327,6 +327,7 @@ def optimize_model(config_path,
     frozen_graph = tf.GraphDef()
     with open(frozen_graph_path, 'rb') as f:
         frozen_graph.ParseFromString(f.read())
+    return frozen_graph

     # apply graph modifications
     if force_nms_cpu:

my config file ssd_inception_v2_coco.json

{
  "source_model": {
    "model_name": "ssd_inception_v2_coco",
    "output_dir": "models"
  },
  "optimization_config": {},
  "benchmark_config": {
    "images_dir": "val2014",
    "annotation_path": "instances_minival2014.json",
    "batch_size": 1,
    "num_images": 100000,
    "output_path": "ssd_inception_v2_coco_trt_fp16.json"
  },
  "assertions": [
    "statistics['map'] > (0.268 - 0.005)"
  ]
}

instances_minival2014.json is from https://dl.dropboxusercontent.com/s/o43o90bna78omob/instances_minival2014.json.zip?dl=0
4. then run
python -m tftrt.examples.object_detection.test ssd_inception_v2_coco.json
5. result
Average Precision (AP) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.278
Average Precision (AP) @[ IoU=0.50 | area= all | maxDets=100 ] = 0.401
Average Precision (AP) @[ IoU=0.75 | area= all | maxDets=100 ] = 0.306
Average Precision (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.027
Average Precision (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.202
Average Precision (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.634
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 1 ] = 0.240
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets= 10 ] = 0.307
Average Recall (AR) @[ IoU=0.50:0.95 | area= all | maxDets=100 ] = 0.308
Average Recall (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.035
Average Recall (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.232
Average Recall (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.695
mkdir: cannot create directory ‘’: No such file or directory
{
"avg_latency_ms": 44.6612452507019,
"avg_throughput_fps": 22.39077738174539,
"map": 0.2776660805925762
}

this result does NOT match the number here:
https://github.com/tensorflow/models/blob/master/research/object_detection/g3doc/detection_model_zoo.md

Is there any official evaluation numbers for this tensorrt project?

attr 'half_pixel_centers' not in Op<name=ResizeBilinear; signature=images:T, size:int32 -> resized_images:float

hello, I open a tensorflow/serving according to this
after I do Convert and serve the model with TF-TRT, finally when I run python /tmp/resnet/resnet_client.py, the tensorflow/serving shows error:

2019-03-14 07:25:19.809001: I tensorflow_serving/model_servers/server.cc:333] Exporting HTTP/REST API at:localhost:8501 ...
2019-03-14 07:25:57.612330: E external/org_tensorflow/tensorflow/core/common_runtime/executor.cc:624] Executor failed to create kernel. Invalid argument: NodeDef mentions attr 'half_pixel_centers' not in Op<name=ResizeBilinear; signature=images:T, size:int32 -> resized_images:float; attr=T:type,allowed=[DT_INT8, DT_UINT8, DT_INT16, DT_UINT16, DT_INT32, DT_INT64, DT_BFLOAT16, DT_HALF, DT_FLOAT, DT_DOUBLE]; attr=align_corners:bool,default=false>; NodeDef: {{node map/while/resize_images/ResizeBilinear}}. (Check whether your GraphDef-interpreting binary is up to date with your GraphDef-generating binary.).
	 [[{{node map/while/resize_images/ResizeBilinear}}]]

how to fix it? anyone can give some advises?

does tf-trt support dynamic shape inputs used for image segmentation models ?

and are there any c++ examples using tf-trt?

AttributeError: module 'tensorflow.contrib.tensorrt' has no attribute 'calib_graph_to_infer_graph' when trying to Calibrate

Hi,

I am trying TensorRT. It works fine as long as I don't try getting the INT8 to work. I got two problems:

1.: is similar to issue #56
2.: occures when trying to calibrate an Resnet v1.5 or SSD-ResNet-1.5. For reproduceability: I used the run_all.sh script from this side.

Maschine:
Ubuntu 18.04, Intel(R) Core(TM) i7-6850K, NVidia 1070, also tried it with the Quadro 5000
Cuda: 10.0
CuDNN: 7
Tensorflow: compiled from source v1.13.1 with TensorRT 5.1.2

When running run_all.sh I get the error:

Traceback (most recent call last):
File "tftrt_sample.py", line 306, in
int8Graph=getINT8InferenceGraph(calibGraph)
File "tftrt_sample.py", line 137, in getINT8InferenceGraph
trt_graph=trt.calib_graph_to_infer_graph(calibGraph)
AttributeError: module 'tensorflow.contrib.tensorrt' has no attribute 'calib_graph_to_infer_graph'

When I curcumvent the line 137 in my own code and don't calibrate the net, the conversion is still working. At least I don't get any errors, but as described in issue #56 the execution for INT8 is almost as slow as for nor converted FP32 and much slower than converted FP32 and FP16.

I tried it with the Quadro 5000 card for both ResNet and SSD-ResNet. With the same outcome and with a compiled Tensorflow v1.14 against TensorRT 5.1.5 same outcome (tested for ResNet) for both Problems.

With kind regards

Conversion with no speed improvement, TRT-TF

My machine info:
Ubuntu 16.04
GeForce GTX 1080
Nvidia Driver Version: 410.78
CUDA Version: 10.0
CUDNN version : 7.5.0
Tensorflow version 1.13.0 from official nvidia docker-hub nvcr.io/nvidia/tensorflow:18.09-py3.
Tensorrt version: from tensorflow package (5.0.2)
Problem description:
I have managed to convert my model to TRT-TF using create_inference_graph method from tensorflow.contrib.tensorrt module. However I do not get statistically significant increase in speed,if any , whatever precision (FP16 or FP32) I use and I am trying to clear up, what can be the reasons for such situation. (Bellow is the console output for optimization)

Questions:

Could it be due to the fact, that tensor shapes are unknown and optimization can not be implemented ?
Is it ok, that my converted .pb file exactly two times larger, than my original frozen not-optimized .pb file ?
Can the few number of supported TensorRT operations be the reason for that ?
I am going to try conversion on Nvidia RTX 2080 Ti , should I expect any benefits in comparison with my current Nvidia GTX 1080 ?

Nvidia GTX1080 console output for optimization :

INFO:tensorflow:Running against TensorRT version 5.0.2
2019-04-11 13:01:33.811488: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:998] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-04-11 13:01:33.811947: I tensorflow/core/grappler/devices.cc:51] Number of eligible GPUs (core count >= 8): 1
2019-04-11 13:01:33.812059: I tensorflow/core/grappler/clusters/single_machine.cc:359] Starting new session
2019-04-11 13:01:33.839792: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 4200000000 Hz
2019-04-11 13:01:33.841161: I tensorflow/compiler/xla/service/service.cc:161] XLA service 0x1e5050c0 executing computations on platform Host. Devices:
2019-04-11 13:01:33.841243: I tensorflow/compiler/xla/service/service.cc:168]   StreamExecutor device (0): <undefined>, <undefined>
2019-04-11 13:01:33.845844: I tensorflow/compiler/xla/service/service.cc:161] XLA service 0x1e52c8e0 executing computations on platform CUDA. Devices:
2019-04-11 13:01:33.845926: I tensorflow/compiler/xla/service/service.cc:168]   StreamExecutor device (0): GeForce GTX 1080, Compute Capability 6.1
2019-04-11 13:01:33.846698: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1433] Found device 0 with properties: 
name: GeForce GTX 1080 major: 6 minor: 1 memoryClockRate(GHz): 1.7335
pciBusID: 0000:01:00.0
totalMemory: 7.93GiB freeMemory: 7.21GiB
2019-04-11 13:01:33.846780: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1512] Adding visible gpu devices: 0
2019-04-11 13:01:34.155614: I tensorflow/core/common_runtime/gpu/gpu_device.cc:984] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-04-11 13:01:34.155652: I tensorflow/core/common_runtime/gpu/gpu_device.cc:990]      0 
2019-04-11 13:01:34.155661: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1003] 0:   N 
2019-04-11 13:01:34.155794: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1115] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 6926 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080, pci bus id: 0000:01:00.0, compute capability: 6.1)
2019-04-11 13:01:34.772860: I tensorflow/contrib/tensorrt/segment/segment.cc:461] There are 7 ops of 4 different types in the graph that are not converted to TensorRT: ConcatV2, Transpose, Placeholder, NoOp, (For more information see https://docs.nvidia.com/deeplearning/dgx/integrate-tf-trt/index.html#support-ops).
2019-04-11 13:01:34.789697: I tensorflow/contrib/tensorrt/convert/convert_graph.cc:928] Number of TensorRT candidate segments: 2
2019-04-11 13:01:34.797911: W tensorflow/contrib/tensorrt/convert/convert_nodes.cc:3728] Validation failed for TensorRTInputPH_0 and input slot 0: Input tensor with shape [?,?,?,3] has an unknown non-batch dimension at dim 1
2019-04-11 13:01:34.797949: W tensorflow/contrib/tensorrt/convert/convert_graph.cc:1036] TensorRT node TRTEngineOp_0 added for segment 0 consisting of 4 nodes failed: Invalid argument: Validation failed for TensorRTInputPH_0 and input slot 0: Input tensor with shape [?,?,?,3] has an unknown non-batch dimension at dim 1. Fallback to TF...
2019-04-11 13:01:34.798309: W tensorflow/contrib/tensorrt/convert/convert_nodes.cc:3728] Validation failed for TensorRTInputPH_0 and input slot 0: Input tensor with shape [?,3,?,?] has an unknown non-batch dimension at dim 2
2019-04-11 13:01:34.798325: W tensorflow/contrib/tensorrt/convert/convert_graph.cc:1036] TensorRT node TRTEngineOp_1 added for segment 1 consisting of 461 nodes failed: Invalid argument: Validation failed for TensorRTInputPH_0 and input slot 0: Input tensor with shape [?,3,?,?] has an unknown non-batch dimension at dim 2. Fallback to TF...
2019-04-11 13:01:34.810444: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:581] Optimization results for grappler item: tf_graph
2019-04-11 13:01:34.810478: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   constant folding: Graph size after: 468 nodes (0), 478 edges (0), time = 61.211ms.
2019-04-11 13:01:34.810485: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   layout: Graph size after: 478 nodes (10), 484 edges (6), time = 25.671ms.
2019-04-11 13:01:34.810490: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   constant folding: Graph size after: 473 nodes (-5), 484 edges (0), time = 40.989ms.
2019-04-11 13:01:34.810495: I tensorflow/core/grappler/optimizers/meta_optimizer.cc:583]   TensorRTOptimizer: Graph size after: 473 nodes (0), 484 edges (0), time = 261.698ms.

download_model function call bug in "tensorrt/tftrt/examples/object_detection/object_detection.py"

In the sample file, "tensorrt/tftrt/examples/object_detection/object_detection.py"

The following line seemed to need to be changed:
261 config_path, checkpoint_path = download_model(
262 model_name=model_name,
263 input_dir=input_dir,
264 output_dir=output_dir)

261 config_path, checkpoint_path = download_model(
262 model_name=model_name,
263 input_dir=None,
264 output_dir=input_dir)

Otherwise, the given sample script for Download would throw out file path error.

TensorRT Mistmatch

OS and version: Ubuntu 18.04
GPU type: NVIDIA GeForce GTX 1080 Ti Founder's Edition
nvidia driver version: 418.56
CUDA version: 10.0
cuDNN version: 7.4.1
Python version: 3.6.7
TensorFlow version:1.13.1
TensorRT version: 5.1.2

Following the instructions per the TensorRT manual, I used the frozen model to implement Tensorflow-RT:

import tensorflow.contrib.tensorrt as trt

trt_graph = trt.create_inference_graph(input_graph_def=frozen_graph, outputs=[out.op.name for out in model.outputs], max_batch_size=1,max_workspace_size_bytes=2 << 20, precision_mode="fp16")
tf.train.write_graph(trt_graph, "model", "tfrt_model.pb", as_text=False)

However, then I get the error:

WARNING:tensorflow:TensorRT mismatch. Compiled against version 5.0.2, but loaded 5.1.2. Things may not work.

What exactly is not lining up? Is my CUDA version too low? Is my cuDNN too low as well? Does it need to be upgraded to 10.1? Or should I downgrade my TensorRT? Am I missing something? Any help would be greatly appreciated.

No TRTEngineOp

after trt.create_inference_graph() , still no "TRTEngineOp", why?

object detection accuracy bug: mAP does not match with the tensorflow released data

we tried to repeate the tf pretrained model accuracy on tensorrt, and we found that the "ssd_inception_v2_coco" accuracy on tensorrt is not the same as the tensorflow released data:(24 vs 27)
could you help to find out why?

see tensorflow release data here

tensorrt ie not enable

I want to use tensorrt optimize my ssd models.but when I use the code like this:

from tftrt.examples.object_detection import optimize_model
import os
model_dir='/home/amax/tensorrt/models/ssd_mobilenet_v1_coco_2018_01_28'
config_path=os.path.join(model_dir,'pipeline.config')
checkpoint_path=os.path.join(model_dir,'model.ckpt')
frozen_graph = optimize_model(
    config_path=config_path,
    checkpoint_path=checkpoint_path,
    use_trt=True,
    precision_mode='FP16'
)

something like this:

Traceback (most recent call last):
  File "optimizer_model.py", line 10, in <module>
    precision_mode='FP16'
  File "/home/amax/tensorrt/tftrt/examples/object_detection/object_detection.py", line 352, in optimize_model
    minimum_segment_size=minimum_segment_size)
  File "/home/amax/anaconda3/lib/python3.5/site-packages/tensorflow/contrib/tensorrt/python/trt_convert.py", line 153, in create_inference_graph
    int(msg[0]))
tensorflow.python.framework.errors_impl.FailedPreconditionError: TensorRT is not enabled!

my env:

tensorflow1.10 source compiled
python 3.5.5(Anaconda 4.2.0)
Linux amax-Super-Server 4.13.0-32-generic #35~16.04.1-Ubuntu SMP Thu Jan 25 10:13:43 UTC 2018 x86_64 x86_64 x86_64 GNU/Linux
tensorrt tar /usr/local/TensorRT-5.0.2.6
cuda9.1 cudnn7.1.4

I have installed tensorrt and install python whl file.I guess compiled from tensorflow source without tensorrt.I have forget it,but i can from tensorflow.contrib import tensorrt.so I don't know where is the problem.I compile with tensorrt some msg like this:

TensorRT libraries found in one the following directories are not compatible with selected cuda and cudnn installations
/usr/local/TensorRT-5.0.2.6
/usr/local/TensorRT-5.0.2.6/lib
/usr/local/TensorRT-5.0.2.6/lib64

Have someone use tensorrt optimizer the frozen.pb file?

No module named compiler.tensorrt

from tftrt.examples.object_detection import download_model
Traceback (most recent call last):
File "", line 1, in
File "tftrt/examples/object_detection/init.py", line 18, in
from .object_detection import download_model, download_dataset, optimize_model, benchmark_model
File "tftrt/examples/object_detection/object_detection.py", line 22, in
from tensorflow.python.compiler.tensorrt import trt_convert as trt
ImportError: No module named compiler.tensorrt

Error with object-detection examples when using batch size > 1 related to image size

@pooyadavoodi @trevor-m When using batch sizes > 1, the latest object-detection code asks to give a fixed image shape but when I provide it with [600,600], it fails with a seg-fault.
This is the error I'm seeing:

Traceback (most recent call last):
  File "/home/mayroy13/anaconda3/envs/trt-py36/lib/python3.6/runpy.py", line 193, in _run_module_as_main
    "__main__", mod_spec)
  File "/home/mayroy13/anaconda3/envs/trt-py36/lib/python3.6/runpy.py", line 85, in _run_code
    exec(code, run_globals)
  File "/home/mayroy13/Mayank/Mayank/test/tensorrt/tftrt/examples/object_detection/test.py", line 105, in <module>
    test(args.test_config_path)
  File "/home/mayroy13/Mayank/Mayank/test/tensorrt/tftrt/examples/object_detection/test.py", line 81, in test
    **test_config['benchmark_config'])
  File "/home/mayroy13/Mayank/Mayank/test/tensorrt/tftrt/examples/object_detection/object_detection.py", line 641, in benchmark_model
    feed_dict={tf_input: batch_images})
  File "/home/mayroy13/anaconda3/envs/trt-py36/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 950, in run
    run_metadata_ptr)
  File "/home/mayroy13/anaconda3/envs/trt-py36/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1149, in _run
    str(subfeed_t.get_shape())))
ValueError: Cannot feed value of shape (104, 600, 600, 3) for Tensor 'image_tensor:0', which has shape '(128, ?, ?, 3)'

The json file I'm using to call the test module with is:

{
  "model_config": {
    "model_name": "ssd_inception_v2_coco",
    "output_dir": "models",
    "override_nms_score_threshold": 0.3,
    "batch_size": 128
  },
  "optimization_config": {
    "use_trt": true,
    "precision_mode": "FP16",
    "output_path":"engine_dir/ssd_inception_v2_b128_fp16.pb",
    "force_nms_cpu": true,
    "replace_relu6": true,
    "remove_assert": true
  },
  "benchmark_config": {
    "images_dir": "coco/val2017",
    "annotation_path": "coco/annotations/instances_val2017.json",
    "batch_size": 128,
    "image_shape":[600,600],
    "num_images": 1000,
    "output_path": "stats/ssd_inception_v2_coco_trt_fp16_b128.json"
  },
  "assertions": [
    "statistics['map'] > (0.268 - 0.005)"
  ]
}

Hoping anyone here can help me sort this error out.

when force_nms_cpu is true and use_trt is true, will be core dumped

Error message:
2019-07-31 20:59:13.742585: I tensorflow/compiler/tf2tensorrt/convert/convert_graph.cc:730] Number of TensorRT candidate segments: 3 2019-07-31 20:59:13.807654: F tensorflow/core/util/device_name_utils.cc:92] Check failed: IsJobName(job) Aborted (core dumped)

Tensorflow version: from 1.13.1 to the latest all have this issue.
TensorRT version:TensorRT-5.0.2.6
cuDNN: cudnn-10.0-linux-x64-v7.4.2.24.tar
OS: ubuntu 16.04
tensorflow/tensorrt version: from r1.13 to the master all have this issue
models: ssd_mobilenet_v1_coco,ssd_resnet50_v1_fpn and others

Change tensorrt/tftrt/examples/object_detection/graph_utils.py:98 as follows will fix this core dumped.
98 #node.device = '/device:CPU:0'
99 node.device = '/job:localhost/replica:0/task:0/device:CPU:0'

Will change this?

How to clear device field in the TF-TRT model

hello, I have trained an inception model, then i use function create_inference_graph() to convert to tf-trt FP16 model.

the strange thing is that, although i set CUDA_VISIBLE_DEVICES=0,1,2, the generated TF-TRT model specifies device field in GraphDef to a single GPU, as the following image shows,

so how could i clear device field in the graphdef, or assign the device field to GPU0,1,2？

Support for XGBoost

XGBoost is a commonly used ML framework. It supports multiple saved model formats like .bst and .joblib. There does not currently exist a performant XGBoost server for inferencing.

It would be great to have TensorRT support for XGBoost models for both CPU and GPU. Unfortunately I don't have any real models other than our toy models. but TensorRT should work on everything XGBoost has to offer.

https://xgboost.readthedocs.io/en/latest/gpu/

Not found: Container TF-TRT-Calibration does not exist

Hi， all；
I am a newbie in tensorflow-trt, when i try to convert a resnet50 model to int8 and calibrate it using imagenet validation data, I meet the following error:

2019-07-17 00:49:44.429295: W tensorflow/core/framework/op_kernel.cc:1622] OP_REQUIRES failed at get_calibration_data_op.cc:48 : Not found: Container TF-TRT-Calibration does not exist. (Could not find resource: TF-TRT-Calibration/TRTEngineOp_0)
Traceback (most recent call last):
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1348, in _do_call
return fn(*args)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1333, in _run_fn
target_list, run_metadata)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1426, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.NotFoundError: 2 root error(s) found.
(0) Not found: Container TF-TRT-Calibration does not exist. (Could not find resource: TF-TRT-Calibration/TRTEngineOp_0)
[[{{node GetCalibrationDataOp}}]]
[[GetCalibrationDataOp/_5]]
(1) Not found: Container TF-TRT-Calibration does not exist. (Could not find resource: TF-TRT-Calibration/TRTEngineOp_0)
[[{{node GetCalibrationDataOp}}]]
0 successful operations.
0 derived errors ignored.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "image_classification.py", line 745, in
max_workspace_size=args.max_workspace_size)
File "image_classification.py", line 617, in get_frozen_graph
input_map_fn=input_map_fn)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/compiler/tensorrt/trt_convert.py", line 358, in calibrate
self.finalize_calibration()
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/compiler/tensorrt/trt_convert.py", line 788, in finalize_calibration
feed_dict={resource_name_input: node.name})
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 941, in run
run_metadata_ptr)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1164, in _run
feed_dict_tensor, options, run_metadata)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1342, in _do_run
run_metadata)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/client/session.py", line 1367, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.NotFoundError: 2 root error(s) found.
(0) Not found: Container TF-TRT-Calibration does not exist. (Could not find resource: TF-TRT-Calibration/TRTEngineOp_0)
[[node GetCalibrationDataOp (defined at /home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/framework/ops.py:1654) ]]
[[GetCalibrationDataOp/_5]]
(1) Not found: Container TF-TRT-Calibration does not exist. (Could not find resource: TF-TRT-Calibration/TRTEngineOp_0)
[[node GetCalibrationDataOp (defined at /home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/framework/ops.py:1654) ]]
0 successful operations.
0 derived errors ignored.

Original stack trace for 'GetCalibrationDataOp':
File "image_classification.py", line 745, in
max_workspace_size=args.max_workspace_size)
File "image_classification.py", line 617, in get_frozen_graph
input_map_fn=input_map_fn)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/compiler/tensorrt/trt_convert.py", line 358, in calibrate
self.finalize_calibration()
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/compiler/tensorrt/trt_convert.py", line 781, in finalize_calibration
gen_trt_ops.get_calibration_data_op(resource_name_input))
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/compiler/tf2tensorrt/ops/gen_trt_ops.py", line 265, in get_calibration_data_op
"GetCalibrationDataOp", resource_name=resource_name, name=name)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/framework/op_def_library.py", line 793, in _apply_op_helper
op_def=op_def)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/util/deprecation.py", line 507, in new_func
return func(*args, **kwargs)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/framework/ops.py", line 3239, in create_op
attrs, op_def, compute_device)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/framework/ops.py", line 3308, in _create_op_internal
op_def=op_def)
File "/home/zhuh/miniconda3/envs/my-tf/lib/python3.6/site-packages/tensorflow_core/python/framework/ops.py", line 1654, in init
self._traceback = tf_stack.extract_stack()

build_model need to be imported in tensorrt/tftrt/examples/object_detection/init.py

It seemed that build_model need to also be imported together with download_model in tensorrt/tftrt/examples/object_detection/init.py, in order to run the object detection sample code.

Otherwise, import Error will be thrown out.

AttributeError: 'module' object has no attribute 'TrtGraphConverter'

when i use example,i got an error,should i update my tf to 14 or 13?

Invalid argument: Unsupported data type encountered in input 0 in mask_rcnn_resnet50_atrous_coco.

Getting the above error on all the RCNN networks which are supposed to have been tested and woked. The full stack trace follows. Is there a to prevent this error ?

2019-08-19 06:47:33.758161: W tensorflow/core/framework/op_kernel.cc:1502] OP_REQUIRES failed at trt_engine_op.cc:293 : Invalid argument: Unsupported data type encountered in input 0
Traceback (most recent call last):
  File "/usr/lib/python2.7/runpy.py", line 174, in _run_module_as_main
    "__main__", fname, loader, pkg_name)
  File "/usr/lib/python2.7/runpy.py", line 72, in _run_code
    exec code in run_globals
  File "build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/test.py", line 105, in <module>
  File "build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/test.py", line 76, in test
  File "build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/object_detection.py", line 453, in optimize_model
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/compiler/tensorrt/trt_convert.py", line 353, in calibrate
    fetches, feed_dict=feed_dict_fn() if feed_dict_fn else None)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 950, in run
    run_metadata_ptr)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1173, in _run
    feed_dict_tensor, options, run_metadata)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1350, in _do_run
    run_metadata)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/client/session.py", line 1370, in _do_call
    raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.InvalidArgumentError: 2 root error(s) found.
  (0) Invalid argument: Unsupported data type encountered in input 0
         [[node FirstStageFeatureExtractor/resnet_v1_50/resnet_v1_50/block3/unit_1/bottleneck_v1/conv2/TRTEngineOp_2239 (defined at build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/object_det$ction.py:453) ]]
         [[BatchMultiClassNonMaxSuppression_1/MultiClassNonMaxSuppression_5/TRTEngineOp_1500/_7959]]
  (1) Invalid argument: Unsupported data type encountered in input 0
         [[node FirstStageFeatureExtractor/resnet_v1_50/resnet_v1_50/block3/unit_1/bottleneck_v1/conv2/TRTEngineOp_2239 (defined at build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/object_det$ction.py:453) ]]
0 successful operations.
0 derived errors ignored.

Original stack trace for u'FirstStageFeatureExtractor/resnet_v1_50/resnet_v1_50/block3/unit_1/bottleneck_v1/conv2/TRTEngineOp_2239':
  File "/usr/lib/python2.7/runpy.py", line 174, in _run_module_as_main
    "__main__", fname, loader, pkg_name)
  File "/usr/lib/python2.7/runpy.py", line 72, in _run_code
    exec code in run_globals
  File "build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/test.py", line 105, in <module>
  File "build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/test.py", line 76, in test
  File "build/bdist.linux-x86_64/egg/tftrt/examples/object_detection/object_detection.py", line 453, in optimize_model
    feed_dict_fn=feed_dict_fn)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/compiler/tensorrt/trt_convert.py", line 347, in calibrate
    name="")
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/util/deprecation.py", line 507, in new_func
    return func(*args, **kwargs)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/importer.py", line 443, in import_graph_def
    _ProcessNewOps(graph)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/importer.py", line 236, in _ProcessNewOps
    for new_op in graph._add_new_tf_operations(compute_devices=False):  # pylint: disable=protected-access
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 3751, in _add_new_tf_operations
    for c_op in c_api_util.new_tf_operations(self)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 3641, in _create_op_from_tf_operation
    ret = Operation(c_op, self)
  File "/usr/local/lib/python2.7/dist-packages/tensorflow/python/framework/ops.py", line 2005, in __init__
    self._traceback = tf_stack.extract_stack()

Some operations not converted, but stated as supported in TF-TRT user guide.

I'm using TF-TRT to convert my model. The conversion is successful, but inference runs slower after conversion. While trying to figure out what goes wrong, I notice that some operations are not converted, which are however stated as supported in TF-TRT user guide.

The corresponding log is: "There are 111 ops of 18 different types in the graph that are not converted to TensorRT: ExpandDims, Reshape, Mul, DataFormatVecPermute, Unpack, MirrorPad, Placeholder, Tile, Cast, NoOp, Elu, Pack, ResizeNearestNeighbor, Shape, StridedSlice, Identity, GreaterEqual, Select, (For more information see https://docs.nvidia.com/deeplearning/dgx/integrate-tf-trt/index.html#support-ops)"

The supported ops section in user guide is: https://docs.nvidia.com/deeplearning/dgx/tf-trt-user-guide/index.html#usingtftrt. For instance, ExpandDims, Reshape, and Mul are clearly stated as supported.

I'm not sure whether this is the reason to the performance decrease, but this is probably an issue. I tried changing the fraction of memory that tensorRT could use, the max batch size, neither of them really helped.

BTW, I'm using TensorFlow-gpu 1.13.1 installed via pip, TensorRT-5.0.2.6, and the GPU is RTX 2080.

i have installed tensorflow and it has method of tf trt

hello, i am starting tf trt . i just need to know that my tensorflow installation gives method of from tensorflow.contrib import tensorrt as tftrt . As now I have converted my graph into tf rt graph. i was worried that do i need to explicitly install tensor rt packages also . or only conda install tensorflow-gpu and above methods from contrib is what we have to do . as for single file inference , i saw only .06 seconds better performance . please let me know

No speed improvements after TF-TRT optimizing

A small tip which may be useful

FP16 or INT8 does improve the inference speed, but not all hardwares support such precision modes.
NVIDIA hardware and which precision modes each hardware supports:
https://docs.nvidia.com/deeplearning/sdk/tensorrt-support-matrix/index.html#hardware-precision-matrix

System information

OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Linux RHEL 7.6
CPU architecture: power8, ppc64le
TensorFlow version (use command below):1.13.1
Python version: Python 2.7.16 :: Anaconda, Inc.
CUDA/cuDNN version: CUDA 10.1, cuDNN 7.6
GPU model and memory: 2 Tesla V100

Result

Dataset: 4952 amount, 300*300 size

Model	Model size	Num Nodes	Batch size	mAP	Latency (ms)	img/sec
FasterRCNN	69M	6931	64	0.7021	342316	15.28
FasterRCNN (TF-TRT)	53M	6456	64	0.7019	334819	15.17
MaskRCNN	78M	7096	32	0.6977	426658	11.67
MaskRCNN (TF-TRT)	53M	6622	32	0.6974	406786	12.17

Abort(core dumped) when optimize model

I'm trying ssd_inception_v2_coco on ubuntu16.04,gtx1080ti.
I have tf-nightly-GPU(1.14.1.dev20190520), CUDA10.0, TensorRT 5.
But always provokes Abort(core dumped). Below is the log:

2019-05-21 17:16:56.863130: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
WARNING: Logging before flag parsing goes to stderr.
W0521 17:16:57.141786 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/slim/nets/inception_resnet_v2.py:373: The name tf.GraphKeys is deprecated. Please use tf.compat.v1.GraphKeys instead.

W0521 17:16:57.168691 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/slim/nets/mobilenet/mobilenet.py:389: The name tf.nn.avg_pool is deprecated. Please use tf.nn.avg_pool2d instead.

/home/zmsunnyday/anaconda3/envs/tf/lib/python3.6/runpy.py:125: RuntimeWarning: 'tftrt.examples.object_detection.test' found in sys.modules after import of package 'tftrt.examples.object_detection', but prior to execution of 'tftrt.examples.object_detection.test'; this may result in unpredictable behaviour
warn(RuntimeWarning(msg))
{
"assertions": [
"statistics['map'] > (0.268 - 0.005)"
],
"benchmark_config": {
"annotation_path": "coco/annotations/instances_val2017.json",
"batch_size": 1,
"image_shape": [
600,
600
],
"images_dir": "coco/val2017",
"num_images": 4096,
"output_path": "stats/ssd_inception_v2_coco_trt_fp16.json"
},
"optimization_config": {
"force_nms_cpu": true,
"max_batch_size": 1,
"override_nms_score_threshold": 0.3,
"precision_mode": "FP16",
"remove_assert": true,
"replace_relu6": true,
"use_trt": true
},
"source_model": {
"model_name": "ssd_inception_v2_coco",
"output_dir": "models"
}
}
--2019-05-21 17:16:57-- http://download.tensorflow.org/models/object_detection/ssd_inception_v2_coco_2018_01_28.tar.gz
Resolving download.tensorflow.org (download.tensorflow.org)... 172.217.163.240, 2404:6800:4008:802::2010
Connecting to download.tensorflow.org (download.tensorflow.org)|172.217.163.240|:80... connected.
HTTP request sent, awaiting response... 200 OK
Length: 278114232 (265M) [application/x-tar]
Saving to: ‘models/ssd_inception_v2_coco_2018_01_28.tar.gz’

models/ssd_inception_v2_coco_2018_01_28.tar.gz 100%[================================================================================================================>] 265.23M 487KB/s in 9m 45s

2019-05-21 17:26:42 (464 KB/s) - ‘models/ssd_inception_v2_coco_2018_01_28.tar.gz’ saved [278114232/278114232]

W0521 17:26:44.701813 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/object_detection.py:312: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.

W0521 17:26:44.702083 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/object_detection.py:316: The name tf.Session is deprecated. Please use tf.compat.v1.Session instead.

2019-05-21 17:26:44.702732: I tensorflow/core/platform/cpu_feature_guard.cc:142] Your CPU supports instructions that this TensorFlow binary was not compiled to use: AVX2 FMA
2019-05-21 17:26:44.707620: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcuda.so.1
2019-05-21 17:26:44.844914: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.851563: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.852517: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x478e4a0 executing computations on platform CUDA. Devices:
2019-05-21 17:26:44.852529: I tensorflow/compiler/xla/service/service.cc:175] StreamExecutor device (0): GeForce GTX 1080 Ti, Compute Capability 6.1
2019-05-21 17:26:44.852533: I tensorflow/compiler/xla/service/service.cc:175] StreamExecutor device (1): GeForce GTX 1080 Ti, Compute Capability 6.1
2019-05-21 17:26:44.873199: I tensorflow/core/platform/profile_utils/cpu_utils.cc:94] CPU Frequency: 3696000000 Hz
2019-05-21 17:26:44.873784: I tensorflow/compiler/xla/service/service.cc:168] XLA service 0x48085f0 executing computations on platform Host. Devices:
2019-05-21 17:26:44.873798: I tensorflow/compiler/xla/service/service.cc:175] StreamExecutor device (0): ,
2019-05-21 17:26:44.874920: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.875697: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:01:00.0
2019-05-21 17:26:44.875733: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.877023: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 1 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:02:00.0
2019-05-21 17:26:44.877035: I tensorflow/stream_executor/platform/default/dlopen_checker.cc:62] Not built with GPU enabled. Skip GPU library dlopen check.
2019-05-21 17:26:44.877072: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.877849: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.878631: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.879400: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:44.880162: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0, 1
2019-05-21 17:26:45.190090: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-05-21 17:26:45.190132: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187] 0 1
2019-05-21 17:26:45.190139: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0: N Y
2019-05-21 17:26:45.190143: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 1: Y N
2019-05-21 17:26:45.190720: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:45.191507: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:45.192280: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:45.193122: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9987 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
2019-05-21 17:26:45.193523: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:45.194319: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 10440 MB memory) -> physical GPU (device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:02:00.0, compute capability: 6.1)
W0521 17:26:45.197109 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:346: The name tf.gfile.MakeDirs is deprecated. Please use tf.io.gfile.makedirs instead.

W0521 17:26:45.197308 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:113: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.

W0521 17:26:45.197858 140549139449600 deprecation.py:323] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:299: to_float (from tensorflow.python.ops.math_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.cast instead.
W0521 17:26:45.198993 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/core/preprocessor.py:2353: The name tf.image.resize_images is deprecated. Please use tf.image.resize instead.

W0521 17:26:45.208121 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/meta_architectures/ssd_meta_arch.py:537: The name tf.variable_scope is deprecated. Please use tf.compat.v1.variable_scope instead.

W0521 17:26:45.209380 140549139449600 deprecation.py:506] From /home/zmsunnyday/anaconda3/envs/tf/lib/python3.6/site-packages/tensorflow/python/ops/init_ops.py:1251: calling VarianceScaling.init (from tensorflow.python.ops.init_ops) with dtype is deprecated and will be removed in a future version.
Instructions for updating:
Call initializer instance with the dtype argument instead of passing it to the constructor
W0521 17:26:47.751603 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/predictors/convolutional_box_predictor.py:147: The name tf.logging.info is deprecated. Please use tf.compat.v1.logging.info instead.

W0521 17:26:47.953805 140549139449600 deprecation.py:323] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/core/post_processing.py:183: add_dispatch_support..wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.where in 2.0, which has the same broadcast rule as np.where
W0521 17:26:49.966654 140549139449600 deprecation.py:323] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:328: get_or_create_global_step (from tensorflow.contrib.framework.python.ops.variables) is deprecated and will be removed in a future version.
Instructions for updating:
Please switch to tf.train.get_or_create_global_step
W0521 17:26:49.968904 140549139449600 deprecation.py:323] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:482: print_model_analysis (from tensorflow.contrib.tfprof.model_analyzer) is deprecated and will be removed after 2018-01-01.
Instructions for updating:
Use tf.profiler.profile(graph, run_meta, op_log, cmd, options). Build options with tf.profiler.ProfileOptionBuilder. See README.md for details
W0521 17:26:49.969716 140549139449600 deprecation.py:323] From /home/zmsunnyday/anaconda3/envs/tf/lib/python3.6/site-packages/tensorflow/python/profiler/internal/flops_registry.py:142: tensor_shape_from_node_def_name (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.compat.v1.graph_util.tensor_shape_from_node_def_name
18 ops no flops stats due to incomplete shapes.
Parsing Inputs...
Incomplete shape.

=========================Options=============================
-max_depth 10000
-min_bytes 0
-min_peak_bytes 0
-min_residual_bytes 0
-min_output_bytes 0
-min_micros 0
-min_accelerator_micros 0
-min_cpu_micros 0
-min_params 0
-min_float_ops 0
-min_occurrence 0
-step -1
-order_by name
-account_type_regexes _trainable_variables
-start_name_regexes .*
-trim_name_regexes .BatchNorm.
-show_name_regexes .*
-hide_name_regexes
-account_displayed_op_only true
-select params
-output stdout:

==================Model Analysis Report======================
Incomplete shape.

Doc:
scope: The nodes in the model graph are organized by their names, which is hierarchical like filesystem.
param: Number of parameters (in the Variable).

Profile:
node name | # parameters
_TFProfRoot (--/25.01m params)
BoxPredictor_0 (--/1.48m params)
BoxPredictor_0/BoxEncodingPredictor (--/62.22k params)
BoxPredictor_0/BoxEncodingPredictor/biases (12, 12/12 params)
BoxPredictor_0/BoxEncodingPredictor/weights (3x3x576x12, 62.21k/62.21k params)
BoxPredictor_0/ClassPredictor (--/1.42m params)
BoxPredictor_0/ClassPredictor/biases (273, 273/273 params)
BoxPredictor_0/ClassPredictor/weights (3x3x576x273, 1.42m/1.42m params)
BoxPredictor_1 (--/5.25m params)
BoxPredictor_1/BoxEncodingPredictor (--/221.21k params)
BoxPredictor_1/BoxEncodingPredictor/biases (24, 24/24 params)
BoxPredictor_1/BoxEncodingPredictor/weights (3x3x1024x24, 221.18k/221.18k params)
BoxPredictor_1/ClassPredictor (--/5.03m params)
BoxPredictor_1/ClassPredictor/biases (546, 546/546 params)
BoxPredictor_1/ClassPredictor/weights (3x3x1024x546, 5.03m/5.03m params)
BoxPredictor_2 (--/2.63m params)
BoxPredictor_2/BoxEncodingPredictor (--/110.62k params)
BoxPredictor_2/BoxEncodingPredictor/biases (24, 24/24 params)
BoxPredictor_2/BoxEncodingPredictor/weights (3x3x512x24, 110.59k/110.59k params)
BoxPredictor_2/ClassPredictor (--/2.52m params)
BoxPredictor_2/ClassPredictor/biases (546, 546/546 params)
BoxPredictor_2/ClassPredictor/weights (3x3x512x546, 2.52m/2.52m params)
BoxPredictor_3 (--/1.31m params)
BoxPredictor_3/BoxEncodingPredictor (--/55.32k params)
BoxPredictor_3/BoxEncodingPredictor/biases (24, 24/24 params)
BoxPredictor_3/BoxEncodingPredictor/weights (3x3x256x24, 55.30k/55.30k params)
BoxPredictor_3/ClassPredictor (--/1.26m params)
BoxPredictor_3/ClassPredictor/biases (546, 546/546 params)
BoxPredictor_3/ClassPredictor/weights (3x3x256x546, 1.26m/1.26m params)
BoxPredictor_4 (--/1.31m params)
BoxPredictor_4/BoxEncodingPredictor (--/55.32k params)
BoxPredictor_4/BoxEncodingPredictor/biases (24, 24/24 params)
BoxPredictor_4/BoxEncodingPredictor/weights (3x3x256x24, 55.30k/55.30k params)
BoxPredictor_4/ClassPredictor (--/1.26m params)
BoxPredictor_4/ClassPredictor/biases (546, 546/546 params)
BoxPredictor_4/ClassPredictor/weights (3x3x256x546, 1.26m/1.26m params)
BoxPredictor_5 (--/657.21k params)
BoxPredictor_5/BoxEncodingPredictor (--/27.67k params)
BoxPredictor_5/BoxEncodingPredictor/biases (24, 24/24 params)
BoxPredictor_5/BoxEncodingPredictor/weights (3x3x128x24, 27.65k/27.65k params)
BoxPredictor_5/ClassPredictor (--/629.54k params)
BoxPredictor_5/ClassPredictor/biases (546, 546/546 params)
BoxPredictor_5/ClassPredictor/weights (3x3x128x546, 628.99k/628.99k params)
FeatureExtractor (--/12.36m params)
FeatureExtractor/InceptionV2 (--/12.36m params)
FeatureExtractor/InceptionV2/Conv2d_1a_7x7 (--/2.71k params)
FeatureExtractor/InceptionV2/Conv2d_1a_7x7/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Conv2d_1a_7x7/depthwise_weights (7x7x3x8, 1.18k/1.18k params)
FeatureExtractor/InceptionV2/Conv2d_1a_7x7/pointwise_weights (1x1x24x64, 1.54k/1.54k params)
FeatureExtractor/InceptionV2/Conv2d_2b_1x1 (--/4.10k params)
FeatureExtractor/InceptionV2/Conv2d_2b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Conv2d_2b_1x1/weights (1x1x64x64, 4.10k/4.10k params)
FeatureExtractor/InceptionV2/Conv2d_2c_3x3 (--/110.59k params)
FeatureExtractor/InceptionV2/Conv2d_2c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Conv2d_2c_3x3/weights (3x3x64x192, 110.59k/110.59k params)
FeatureExtractor/InceptionV2/Mixed_3b (--/218.11k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_0 (--/12.29k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_0/Conv2d_0a_1x1 (--/12.29k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_0/Conv2d_0a_1x1/weights (1x1x192x64, 12.29k/12.29k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_1 (--/49.15k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_1/Conv2d_0a_1x1 (--/12.29k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_1/Conv2d_0a_1x1/weights (1x1x192x64, 12.29k/12.29k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_1/Conv2d_0b_3x3 (--/36.86k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_1/Conv2d_0b_3x3/weights (3x3x64x64, 36.86k/36.86k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2 (--/150.53k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0a_1x1 (--/12.29k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0a_1x1/weights (1x1x192x64, 12.29k/12.29k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0b_3x3 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0b_3x3/weights (3x3x64x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0c_3x3 (--/82.94k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_2/Conv2d_0c_3x3/weights (3x3x96x96, 82.94k/82.94k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_3 (--/6.14k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_3/Conv2d_0b_1x1 (--/6.14k params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3b/Branch_3/Conv2d_0b_1x1/weights (1x1x192x32, 6.14k/6.14k params)
FeatureExtractor/InceptionV2/Mixed_3c (--/259.07k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_0 (--/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_0/Conv2d_0a_1x1 (--/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_0/Conv2d_0a_1x1/weights (1x1x256x64, 16.38k/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_1 (--/71.68k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_1/Conv2d_0a_1x1 (--/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_1/Conv2d_0a_1x1/weights (1x1x256x64, 16.38k/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_1/Conv2d_0b_3x3 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_1/Conv2d_0b_3x3/weights (3x3x64x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2 (--/154.62k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0a_1x1 (--/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0a_1x1/weights (1x1x256x64, 16.38k/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0b_3x3 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0b_3x3/weights (3x3x64x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0c_3x3 (--/82.94k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_2/Conv2d_0c_3x3/weights (3x3x96x96, 82.94k/82.94k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_3 (--/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_3/Conv2d_0b_1x1 (--/16.38k params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_3c/Branch_3/Conv2d_0b_1x1/weights (1x1x256x64, 16.38k/16.38k params)
FeatureExtractor/InceptionV2/Mixed_4a (--/384.00k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_0 (--/225.28k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_0/Conv2d_0a_1x1 (--/40.96k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_0/Conv2d_0a_1x1/weights (1x1x320x128, 40.96k/40.96k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_0/Conv2d_1a_3x3 (--/184.32k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_0/Conv2d_1a_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_0/Conv2d_1a_3x3/weights (3x3x128x160, 184.32k/184.32k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1 (--/158.72k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_0a_1x1 (--/20.48k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_0a_1x1/weights (1x1x320x64, 20.48k/20.48k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_0b_3x3 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_0b_3x3/weights (3x3x64x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_1a_3x3 (--/82.94k params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_1a_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4a/Branch_1/Conv2d_1a_3x3/weights (3x3x96x96, 82.94k/82.94k params)
FeatureExtractor/InceptionV2/Mixed_4b (--/608.26k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_0 (--/129.02k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_0/Conv2d_0a_1x1 (--/129.02k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_0/Conv2d_0a_1x1/weights (1x1x576x224, 129.02k/129.02k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_1 (--/92.16k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_1/Conv2d_0a_1x1 (--/36.86k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_1/Conv2d_0a_1x1/weights (1x1x576x64, 36.86k/36.86k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_1/Conv2d_0b_3x3 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_1/Conv2d_0b_3x3/weights (3x3x64x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2 (--/313.34k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0a_1x1 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0a_1x1/weights (1x1x576x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0b_3x3 (--/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0b_3x3/weights (3x3x96x128, 110.59k/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0c_3x3 (--/147.46k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_2/Conv2d_0c_3x3/weights (3x3x128x128, 147.46k/147.46k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_3 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_3/Conv2d_0b_1x1 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4b/Branch_3/Conv2d_0b_1x1/weights (1x1x576x128, 73.73k/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4c (--/663.55k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_0 (--/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_0/Conv2d_0a_1x1 (--/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_0/Conv2d_0a_1x1/weights (1x1x576x192, 110.59k/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_1 (--/165.89k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_1/Conv2d_0a_1x1 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_1/Conv2d_0a_1x1/weights (1x1x576x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_1/Conv2d_0b_3x3 (--/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_1/Conv2d_0b_3x3/weights (3x3x96x128, 110.59k/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2 (--/313.34k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0a_1x1 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0a_1x1/weights (1x1x576x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0b_3x3 (--/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0b_3x3/weights (3x3x96x128, 110.59k/110.59k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0c_3x3 (--/147.46k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_2/Conv2d_0c_3x3/weights (3x3x128x128, 147.46k/147.46k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_3 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_3/Conv2d_0b_1x1 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4c/Branch_3/Conv2d_0b_1x1/weights (1x1x576x128, 73.73k/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4d (--/893.95k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_0 (--/92.16k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_0/Conv2d_0a_1x1 (--/92.16k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_0/Conv2d_0a_1x1/weights (1x1x576x160, 92.16k/92.16k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_1 (--/258.05k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_1/Conv2d_0a_1x1 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_1/Conv2d_0a_1x1/weights (1x1x576x128, 73.73k/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_1/Conv2d_0b_3x3 (--/184.32k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_1/Conv2d_0b_3x3/weights (3x3x128x160, 184.32k/184.32k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2 (--/488.45k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0a_1x1 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0a_1x1/weights (1x1x576x128, 73.73k/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0b_3x3 (--/184.32k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0b_3x3/weights (3x3x128x160, 184.32k/184.32k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0c_3x3 (--/230.40k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_2/Conv2d_0c_3x3/weights (3x3x160x160, 230.40k/230.40k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_3 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_3/Conv2d_0b_1x1 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4d/Branch_3/Conv2d_0b_1x1/weights (1x1x576x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4e (--/1.11m params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_0 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_0/Conv2d_0a_1x1 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_0/Conv2d_0a_1x1/weights (1x1x576x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_1 (--/294.91k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_1/Conv2d_0a_1x1 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_1/Conv2d_0a_1x1/weights (1x1x576x128, 73.73k/73.73k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_1/Conv2d_0b_3x3 (--/221.18k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_1/Conv2d_0b_3x3/weights (3x3x128x192, 221.18k/221.18k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2 (--/700.42k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0a_1x1 (--/92.16k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0a_1x1/weights (1x1x576x160, 92.16k/92.16k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0b_3x3 (--/276.48k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0b_3x3/weights (3x3x160x192, 276.48k/276.48k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0c_3x3 (--/331.78k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_2/Conv2d_0c_3x3/weights (3x3x192x192, 331.78k/331.78k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_3 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_3/Conv2d_0b_1x1 (--/55.30k params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_4e/Branch_3/Conv2d_0b_1x1/weights (1x1x576x96, 55.30k/55.30k params)
FeatureExtractor/InceptionV2/Mixed_5a (--/1.44m params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_0 (--/294.91k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_0/Conv2d_0a_1x1 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_0/Conv2d_0a_1x1/weights (1x1x576x128, 73.73k/73.73k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_0/Conv2d_1a_3x3 (--/221.18k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_0/Conv2d_1a_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_0/Conv2d_1a_3x3/weights (3x3x128x192, 221.18k/221.18k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1 (--/1.14m params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_0a_1x1 (--/110.59k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_0a_1x1/weights (1x1x576x192, 110.59k/110.59k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_0b_3x3 (--/442.37k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_0b_3x3/weights (3x3x192x256, 442.37k/442.37k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_1a_3x3 (--/589.82k params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_1a_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5a/Branch_1/Conv2d_1a_3x3/weights (3x3x256x256, 589.82k/589.82k params)
FeatureExtractor/InceptionV2/Mixed_5b (--/2.18m params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_0 (--/360.45k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_0/Conv2d_0a_1x1 (--/360.45k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_0/Conv2d_0a_1x1/weights (1x1x1024x352, 360.45k/360.45k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_1 (--/749.57k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_1/Conv2d_0a_1x1 (--/196.61k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_1/Conv2d_0a_1x1/weights (1x1x1024x192, 196.61k/196.61k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_1/Conv2d_0b_3x3 (--/552.96k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_1/Conv2d_0b_3x3/weights (3x3x192x320, 552.96k/552.96k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2 (--/937.98k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0a_1x1 (--/163.84k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0a_1x1/weights (1x1x1024x160, 163.84k/163.84k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0b_3x3 (--/322.56k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0b_3x3/weights (3x3x160x224, 322.56k/322.56k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0c_3x3 (--/451.58k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_2/Conv2d_0c_3x3/weights (3x3x224x224, 451.58k/451.58k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_3 (--/131.07k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_3/Conv2d_0b_1x1 (--/131.07k params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5b/Branch_3/Conv2d_0b_1x1/weights (1x1x1024x128, 131.07k/131.07k params)
FeatureExtractor/InceptionV2/Mixed_5c (--/2.28m params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_0 (--/360.45k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_0/Conv2d_0a_1x1 (--/360.45k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_0/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_0/Conv2d_0a_1x1/weights (1x1x1024x352, 360.45k/360.45k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_1 (--/749.57k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_1/Conv2d_0a_1x1 (--/196.61k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_1/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_1/Conv2d_0a_1x1/weights (1x1x1024x192, 196.61k/196.61k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_1/Conv2d_0b_3x3 (--/552.96k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_1/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_1/Conv2d_0b_3x3/weights (3x3x192x320, 552.96k/552.96k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2 (--/1.04m params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0a_1x1 (--/196.61k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0a_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0a_1x1/weights (1x1x1024x192, 196.61k/196.61k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0b_3x3 (--/387.07k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0b_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0b_3x3/weights (3x3x192x224, 387.07k/387.07k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0c_3x3 (--/451.58k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0c_3x3/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_2/Conv2d_0c_3x3/weights (3x3x224x224, 451.58k/451.58k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_3 (--/131.07k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_3/Conv2d_0b_1x1 (--/131.07k params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_3/Conv2d_0b_1x1/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c/Branch_3/Conv2d_0b_1x1/weights (1x1x1024x128, 131.07k/131.07k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_2_1x1_256 (--/262.14k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_2_1x1_256/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_2_1x1_256/weights (1x1x1024x256, 262.14k/262.14k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_3_1x1_128 (--/65.54k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_3_1x1_128/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_3_1x1_128/weights (1x1x512x128, 65.54k/65.54k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_4_1x1_128 (--/32.77k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_4_1x1_128/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_4_1x1_128/weights (1x1x256x128, 32.77k/32.77k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_5_1x1_64 (--/16.38k params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_5_1x1_64/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_5_1x1_64/weights (1x1x256x64, 16.38k/16.38k params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_2_3x3_s2_512 (--/1.18m params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_2_3x3_s2_512/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_2_3x3_s2_512/weights (3x3x256x512, 1.18m/1.18m params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_3_3x3_s2_256 (--/294.91k params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_3_3x3_s2_256/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_3_3x3_s2_256/weights (3x3x128x256, 294.91k/294.91k params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_4_3x3_s2_256 (--/294.91k params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_4_3x3_s2_256/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_4_3x3_s2_256/weights (3x3x128x256, 294.91k/294.91k params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_5_3x3_s2_128 (--/73.73k params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_5_3x3_s2_128/BatchNorm (--/0 params)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_5_3x3_s2_128/weights (3x3x64x128, 73.73k/73.73k params)

======================End of Report==========================
18 ops no flops stats due to incomplete shapes.
Parsing Inputs...
Incomplete shape.

=========================Options=============================
-max_depth 10000
-min_bytes 0
-min_peak_bytes 0
-min_residual_bytes 0
-min_output_bytes 0
-min_micros 0
-min_accelerator_micros 0
-min_cpu_micros 0
-min_params 0
-min_float_ops 1
-min_occurrence 0
-step -1
-order_by float_ops
-account_type_regexes .*
-start_name_regexes .*
-trim_name_regexes .BatchNorm.,.Initializer.,.Regularizer.,.BiasAdd.
-show_name_regexes .*
-hide_name_regexes
-account_displayed_op_only true
-select float_ops
-output stdout:

==================Model Analysis Report======================
Incomplete shape.

Doc:
scope: The nodes in the model graph are organized by their names, which is hierarchical like filesystem.
flops: Number of float operations. Note: Please read the implementation for the math behind it.

Profile:
node name | # float_ops
_TFProfRoot (--/9.62b flops)
FeatureExtractor/InceptionV2/InceptionV2/Conv2d_2c_3x3/Conv2D (1.24b/1.24b flops)
BoxPredictor_0/ClassPredictor/Conv2D (1.02b/1.02b flops)
BoxPredictor_1/ClassPredictor/Conv2D (1.01b/1.01b flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5a/Branch_1/Conv2d_0b_3x3/Conv2D (319.39m/319.39m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_2/Conv2d_0c_3x3/Conv2D (239.54m/239.54m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_2/Conv2d_0c_3x3/Conv2D (239.54m/239.54m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_2/Conv2d_0c_3x3/Conv2D (239.54m/239.54m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_2/Conv2d_0b_3x3/Conv2D (199.62m/199.62m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_2/Conv2d_0c_3x3/Conv2D (166.35m/166.35m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_1/Conv2d_0b_3x3/Conv2D (159.69m/159.69m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4a/Branch_1/Conv2d_0b_3x3/Conv2D (159.69m/159.69m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_1/Conv2d_0b_3x3/Conv2D (159.69m/159.69m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_2/Conv2d_0b_3x3/Conv2D (159.69m/159.69m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_2/Conv2d_0b_3x3/Conv2D (159.69m/159.69m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_2/Conv2d_0b_3x3/Conv2D (133.08m/133.08m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4a/Branch_0/Conv2d_1a_3x3/Conv2D (133.08m/133.08m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_1/Conv2d_0b_3x3/Conv2D (133.08m/133.08m flops)
BoxPredictor_2/ClassPredictor/Conv2D (125.80m/125.80m flops)
FeatureExtractor/InceptionV2/InceptionV2/Conv2d_1a_7x7/separable_conv2d (69.12m/122.04m flops)
FeatureExtractor/InceptionV2/InceptionV2/Conv2d_1a_7x7/separable_conv2d/depthwise (52.92m/52.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4a/Branch_0/Conv2d_0a_1x1/Conv2D (118.29m/118.29m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5a/Branch_1/Conv2d_1a_3x3/Conv2D (117.96m/117.96m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_1/Conv2d_0b_3x3/Conv2D (110.59m/110.59m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_1/Conv2d_0b_3x3/Conv2D (110.59m/110.59m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_2/Conv2d_0c_3x3/Conv2D (106.46m/106.46m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_2/Conv2d_0c_3x3/Conv2D (106.46m/106.46m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_1/Conv2d_0b_3x3/Conv2D (106.46m/106.46m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_0/Conv2d_0a_1x1/Conv2D (93.16m/93.16m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_2/Conv2d_0c_3x3/Conv2D (90.32m/90.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_2/Conv2d_0c_3x3/Conv2D (90.32m/90.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_1/Conv2d_0b_3x3/Conv2D (79.85m/79.85m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_2/Conv2d_0b_3x3/Conv2D (79.85m/79.85m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_2/Conv2d_0b_3x3/Conv2D (79.85m/79.85m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_0/Conv2d_0a_1x1/Conv2D (79.85m/79.85m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5a/Branch_1/Conv2d_0a_1x1/Conv2D (79.85m/79.85m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_2/Conv2d_0b_3x3/Conv2D (77.41m/77.41m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_0/Conv2d_0a_1x1/Conv2D (72.09m/72.09m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_0/Conv2d_0a_1x1/Conv2D (72.09m/72.09m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_2/Conv2d_0a_1x1/Conv2D (66.54m/66.54m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_0/Conv2d_0a_1x1/Conv2D (66.54m/66.54m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_2/Conv2d_0b_3x3/Conv2D (64.51m/64.51m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4a/Branch_1/Conv2d_1a_3x3/Conv2D (59.89m/59.89m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4a/Branch_1/Conv2d_0a_1x1/Conv2D (59.15m/59.15m flops)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_2_3x3_s2_512/Conv2D (58.98m/58.98m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_1/Conv2d_0a_1x1/Conv2D (53.23m/53.23m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_3/Conv2d_0b_1x1/Conv2D (53.23m/53.23m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_3/Conv2d_0b_1x1/Conv2D (53.23m/53.23m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_1/Conv2d_0a_1x1/Conv2D (53.23m/53.23m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5a/Branch_0/Conv2d_0a_1x1/Conv2D (53.23m/53.23m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_2/Conv2d_0a_1x1/Conv2D (53.23m/53.23m flops)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_2_1x1_256/Conv2D (52.43m/52.43m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_3/Conv2d_0b_1x1/Conv2D (47.32m/47.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_2/Conv2d_0a_1x1/Conv2D (47.32m/47.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_1/Conv2d_0a_1x1/Conv2D (47.32m/47.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_0/Conv2d_0a_1x1/Conv2D (47.32m/47.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Conv2d_2b_1x1/Conv2D (46.08m/46.08m flops)
BoxPredictor_0/BoxEncodingPredictor/Conv2D (44.91m/44.91m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5a/Branch_0/Conv2d_1a_3x3/Conv2D (44.24m/44.24m flops)
BoxPredictor_1/BoxEncodingPredictor/Conv2D (44.24m/44.24m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_3/Conv2d_0b_1x1/Conv2D (39.92m/39.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_3/Conv2d_0b_1x1/Conv2D (39.92m/39.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_2/Conv2d_0a_1x1/Conv2D (39.92m/39.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_0/Conv2d_0a_1x1/Conv2D (39.92m/39.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_1/Conv2d_0a_1x1/Conv2D (39.92m/39.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_2/Conv2d_0a_1x1/Conv2D (39.92m/39.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_1/Conv2d_0b_3x3/Conv2D (39.92m/39.92m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_1/Conv2d_0a_1x1/Conv2D (39.32m/39.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_1/Conv2d_0a_1x1/Conv2D (39.32m/39.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_2/Conv2d_0a_1x1/Conv2D (39.32m/39.32m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_1/Conv2d_0a_1x1/Conv2D (35.49m/35.49m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_0/Conv2d_0a_1x1/Conv2D (35.49m/35.49m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_2/Conv2d_0a_1x1/Conv2D (35.49m/35.49m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_2/Conv2d_0a_1x1/Conv2D (32.77m/32.77m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_1/Conv2d_0a_1x1/Conv2D (26.62m/26.62m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_3/Conv2d_0b_1x1/Conv2D (26.21m/26.21m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_3/Conv2d_0b_1x1/Conv2D (26.21m/26.21m flops)
BoxPredictor_3/ClassPredictor/Conv2D (22.64m/22.64m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_3/Conv2d_0b_1x1/Conv2D (17.74m/17.74m flops)
BoxPredictor_4/ClassPredictor/Conv2D (10.06m/10.06m flops)
BoxPredictor_2/BoxEncodingPredictor/Conv2D (5.53m/5.53m flops)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_3_3x3_s2_256/Conv2D (5.31m/5.31m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3c/Branch_3/AvgPool_0a_3x3/AvgPool (3.33m/3.33m flops)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_3_1x1_128/Conv2D (3.28m/3.28m flops)
FeatureExtractor/InceptionV2/InceptionV2/MaxPool_2a_3x3/MaxPool (3.24m/3.24m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_3b/Branch_3/AvgPool_0a_3x3/AvgPool (2.50m/2.50m flops)
FeatureExtractor/InceptionV2/InceptionV2/MaxPool_3a_3x3/MaxPool (2.50m/2.50m flops)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_4_3x3_s2_256/Conv2D (2.36m/2.36m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4b/Branch_3/AvgPool_0a_3x3/AvgPool (1.87m/1.87m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4c/Branch_3/AvgPool_0a_3x3/AvgPool (1.87m/1.87m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4d/Branch_3/AvgPool_0a_3x3/AvgPool (1.87m/1.87m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4e/Branch_3/AvgPool_0a_3x3/AvgPool (1.87m/1.87m flops)
BoxPredictor_5/ClassPredictor/Conv2D (1.26m/1.26m flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_4a/Branch_2/MaxPool_1a_3x3/MaxPool (1.04m/1.04m flops)
BoxPredictor_3/BoxEncodingPredictor/Conv2D (995.33k/995.33k flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5b/Branch_3/AvgPool_0a_3x3/AvgPool (921.60k/921.60k flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5c/Branch_3/MaxPool_0a_3x3/MaxPool (921.60k/921.60k flops)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_4_1x1_128/Conv2D (589.82k/589.82k flops)
FeatureExtractor/InceptionV2/InceptionV2/Mixed_5a/Branch_2/MaxPool_1a_3x3/MaxPool (518.40k/518.40k flops)
BoxPredictor_4/BoxEncodingPredictor/Conv2D (442.37k/442.37k flops)
Preprocessor/mul (270.00k/270.00k flops)
Preprocessor/sub (270.00k/270.00k flops)
Postprocessor/scale_logits (174.45k/174.45k flops)
FeatureExtractor/InceptionV2/Mixed_5c_2_Conv2d_5_3x3_s2_128/Conv2D (147.46k/147.46k flops)
FeatureExtractor/InceptionV2/Mixed_5c_1_Conv2d_5_1x1_64/Conv2D (131.07k/131.07k flops)
BoxPredictor_0/ClassPredictor/BiasAdd (98.55k/98.55k flops)
BoxPredictor_5/BoxEncodingPredictor/Conv2D (55.30k/55.30k flops)
BoxPredictor_1/ClassPredictor/BiasAdd (54.60k/54.60k flops)
BoxPredictor_2/ClassPredictor/BiasAdd (13.65k/13.65k flops)
BoxPredictor_3/ClassPredictor/BiasAdd (4.91k/4.91k flops)
BoxPredictor_0/BoxEncodingPredictor/BiasAdd (4.33k/4.33k flops)
BoxPredictor_1/BoxEncodingPredictor/BiasAdd (2.40k/2.40k flops)
BoxPredictor_4/ClassPredictor/BiasAdd (2.18k/2.18k flops)
MultipleGridAnchorGenerator/sub (2.17k/2.17k flops)
MultipleGridAnchorGenerator/add_2 (2.17k/2.17k flops)
MultipleGridAnchorGenerator/mul_20 (2.17k/2.17k flops)
MultipleGridAnchorGenerator/mul_19 (2.17k/2.17k flops)
Postprocessor/Decode/truediv_6 (1.92k/1.92k flops)
Postprocessor/Decode/add (1.92k/1.92k flops)
Postprocessor/Decode/add_1 (1.92k/1.92k flops)
Postprocessor/Decode/add_2 (1.92k/1.92k flops)
Postprocessor/Decode/add_3 (1.92k/1.92k flops)
Postprocessor/Decode/get_center_coordinates_and_sizes/add (1.92k/1.92k flops)
Postprocessor/Decode/get_center_coordinates_and_sizes/add_1 (1.92k/1.92k flops)
Postprocessor/Decode/get_center_coordinates_and_sizes/sub (1.92k/1.92k flops)
Postprocessor/Decode/get_center_coordinates_and_sizes/sub_1 (1.92k/1.92k flops)
Postprocessor/Decode/get_center_coordinates_and_sizes/truediv (1.92k/1.92k flops)
Postprocessor/Decode/get_center_coordinates_and_sizes/truediv_1 (1.92k/1.92k flops)
Postprocessor/Decode/mul (1.92k/1.92k flops)
Postprocessor/Decode/mul_1 (1.92k/1.92k flops)
Postprocessor/Decode/mul_3 (1.92k/1.92k flops)
Postprocessor/Decode/truediv_7 (1.92k/1.92k flops)
Postprocessor/Decode/truediv_5 (1.92k/1.92k flops)
Postprocessor/Decode/truediv_4 (1.92k/1.92k flops)
Postprocessor/Decode/truediv_3 (1.92k/1.92k flops)
Postprocessor/Decode/truediv_2 (1.92k/1.92k flops)
Postprocessor/Decode/truediv_1 (1.92k/1.92k flops)
Postprocessor/Decode/truediv (1.92k/1.92k flops)
Postprocessor/Decode/sub_1 (1.92k/1.92k flops)
Postprocessor/Decode/sub (1.92k/1.92k flops)
Postprocessor/Decode/mul_2 (1.92k/1.92k flops)
MultipleGridAnchorGenerator/add_5 (1.20k/1.20k flops)
MultipleGridAnchorGenerator/sub_1 (1.20k/1.20k flops)
MultipleGridAnchorGenerator/mul_28 (1.20k/1.20k flops)
MultipleGridAnchorGenerator/mul_27 (1.20k/1.20k flops)
MultipleGridAnchorGenerator/mul_21 (1.08k/1.08k flops)
BoxPredictor_2/BoxEncodingPredictor/BiasAdd (600/600 flops)
MultipleGridAnchorGenerator/mul_29 (600/600 flops)
BoxPredictor_5/ClassPredictor/BiasAdd (546/546 flops)
MultipleGridAnchorGenerator/mul_36 (300/300 flops)
MultipleGridAnchorGenerator/mul_35 (300/300 flops)
MultipleGridAnchorGenerator/sub_2 (300/300 flops)
MultipleGridAnchorGenerator/add_8 (300/300 flops)
BoxPredictor_3/BoxEncodingPredictor/BiasAdd (216/216 flops)
MultipleGridAnchorGenerator/mul_37 (150/150 flops)
MultipleGridAnchorGenerator/mul_43 (108/108 flops)
MultipleGridAnchorGenerator/mul_44 (108/108 flops)
MultipleGridAnchorGenerator/add_11 (108/108 flops)
MultipleGridAnchorGenerator/sub_3 (108/108 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_41 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_57 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_53 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_51 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_5 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_49 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_47 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_45 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_43 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_23 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_39 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_37 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_35 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_33 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_31 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_3 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_29 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_27 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_25 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_77 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_99 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_97 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_95 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_93 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_91 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_9 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_89 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_87 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_85 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_83 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_81 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_79 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_55 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_75 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_73 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_71 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_7 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_69 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_67 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_65 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_63 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_61 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_59 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_19 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_119 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_139 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_137 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_135 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_133 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_131 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_13 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_129 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_127 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_125 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_123 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_121 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_141 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_117 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_115 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_113 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_111 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_11 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_109 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_107 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_105 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_103 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_101 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_1 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_161 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_11 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_179 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_177 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_175 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_173 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_171 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_17 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_169 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_167 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_165 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_163 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_21 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_159 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_157 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_155 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_153 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_151 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_15 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_149 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_147 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_145 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_143 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_6 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_7 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_69 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_68 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_67 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_66 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_65 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_64 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_63 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_62 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_61 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_60 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_70 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_59 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_58 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_57 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_56 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_55 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_54 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_53 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_52 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_51 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_50 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_89 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_83 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_85 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_86 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_82 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_81 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_80 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_8 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_79 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_87 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_88 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_5 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_78 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_77 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_76 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_75 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_74 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_73 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_9 (100/100 flops)
add (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_72 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_71 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_19 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_29 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_28 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_27 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_26 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_25 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_24 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_23 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_22 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_21 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_20 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_2 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_3 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_18 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_17 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_16 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_15 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_14 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_13 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_12 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_84 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_10 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_1 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_4 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_49 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_48 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_47 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_46 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_45 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_44 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_43 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_42 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_41 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_40 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_39 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_38 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_37 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_36 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_35 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_34 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_33 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_32 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_31 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/mul_30 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_59 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_69 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_68 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_67 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_66 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_65 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_64 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_63 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_62 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_61 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_60 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_6 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_7 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_58 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_57 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_56 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_55 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_54 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_53 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_52 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_51 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_50 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_5 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_8 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_9 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_89 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_88 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_87 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_86 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_85 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_84 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_83 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_82 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_81 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_80 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_49 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_79 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_78 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_77 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_76 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_75 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_74 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_73 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_72 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_71 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_70 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_18 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_28 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_27 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_26 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_25 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_24 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_23 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_22 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_21 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_20 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_2 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_19 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_29 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_17 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_16 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_15 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_14 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_13 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_12 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_11 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_10 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_1 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_39 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_48 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_47 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_46 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_45 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_44 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_43 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_42 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_41 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_40 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_4 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_38 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_37 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_36 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_35 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_34 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_33 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_32 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_31 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_30 (100/100 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Less_3 (100/100 flops)
BoxPredictor_4/BoxEncodingPredictor/BiasAdd (96/96 flops)
MultipleGridAnchorGenerator/mul_45 (54/54 flops)
MultipleGridAnchorGenerator/sub_4 (48/48 flops)
MultipleGridAnchorGenerator/mul_51 (48/48 flops)
MultipleGridAnchorGenerator/mul_52 (48/48 flops)
MultipleGridAnchorGenerator/add_14 (48/48 flops)
BoxPredictor_5/BoxEncodingPredictor/BiasAdd (24/24 flops)
MultipleGridAnchorGenerator/mul_53 (24/24 flops)
MultipleGridAnchorGenerator/mul_17 (19/19 flops)
MultipleGridAnchorGenerator/add (19/19 flops)
MultipleGridAnchorGenerator/add_1 (19/19 flops)
MultipleGridAnchorGenerator/mul_18 (19/19 flops)
MultipleGridAnchorGenerator/mul_60 (12/12 flops)
MultipleGridAnchorGenerator/mul_59 (12/12 flops)
MultipleGridAnchorGenerator/add_17 (12/12 flops)
MultipleGridAnchorGenerator/sub_5 (12/12 flops)
MultipleGridAnchorGenerator/mul_26 (10/10 flops)
MultipleGridAnchorGenerator/add_4 (10/10 flops)
MultipleGridAnchorGenerator/add_3 (10/10 flops)
MultipleGridAnchorGenerator/mul_25 (10/10 flops)
MultipleGridAnchorGenerator/mul_31 (6/6 flops)
MultipleGridAnchorGenerator/mul_54 (6/6 flops)
MultipleGridAnchorGenerator/mul_32 (6/6 flops)
MultipleGridAnchorGenerator/mul_48 (6/6 flops)
MultipleGridAnchorGenerator/mul_56 (6/6 flops)
MultipleGridAnchorGenerator/mul_55 (6/6 flops)
MultipleGridAnchorGenerator/mul_38 (6/6 flops)
MultipleGridAnchorGenerator/mul_39 (6/6 flops)
MultipleGridAnchorGenerator/mul_61 (6/6 flops)
MultipleGridAnchorGenerator/mul_30 (6/6 flops)
MultipleGridAnchorGenerator/mul_47 (6/6 flops)
MultipleGridAnchorGenerator/mul_46 (6/6 flops)
MultipleGridAnchorGenerator/mul_24 (6/6 flops)
MultipleGridAnchorGenerator/truediv_19 (6/6 flops)
MultipleGridAnchorGenerator/truediv_18 (6/6 flops)
MultipleGridAnchorGenerator/truediv_17 (6/6 flops)
MultipleGridAnchorGenerator/truediv_16 (6/6 flops)
MultipleGridAnchorGenerator/truediv_15 (6/6 flops)
MultipleGridAnchorGenerator/mul_23 (6/6 flops)
MultipleGridAnchorGenerator/mul_22 (6/6 flops)
MultipleGridAnchorGenerator/mul_40 (6/6 flops)
MultipleGridAnchorGenerator/add_6 (5/5 flops)
MultipleGridAnchorGenerator/mul_34 (5/5 flops)
MultipleGridAnchorGenerator/mul_33 (5/5 flops)
MultipleGridAnchorGenerator/add_7 (5/5 flops)
MultipleGridAnchorGenerator/mul_14 (3/3 flops)
MultipleGridAnchorGenerator/mul_16 (3/3 flops)
MultipleGridAnchorGenerator/mul_15 (3/3 flops)
MultipleGridAnchorGenerator/mul_42 (3/3 flops)
MultipleGridAnchorGenerator/truediv_14 (3/3 flops)
MultipleGridAnchorGenerator/add_10 (3/3 flops)
MultipleGridAnchorGenerator/add_9 (3/3 flops)
MultipleGridAnchorGenerator/mul_41 (3/3 flops)
MultipleGridAnchorGenerator/mul_50 (2/2 flops)
MultipleGridAnchorGenerator/mul_49 (2/2 flops)
MultipleGridAnchorGenerator/add_12 (2/2 flops)
MultipleGridAnchorGenerator/add_13 (2/2 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_19 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_26 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_40 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_2 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_27 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_29 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_20 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_21 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_28 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_22 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_23 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_24 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_25 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_3 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_30 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_31 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_32 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_33 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_34 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_35 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_36 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_37 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_38 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_39 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_4 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_47 (1/1 flops)
MultipleGridAnchorGenerator/mul_13 (1/1 flops)
MultipleGridAnchorGenerator/truediv_1 (1/1 flops)
MultipleGridAnchorGenerator/truediv (1/1 flops)
MultipleGridAnchorGenerator/mul_9 (1/1 flops)
MultipleGridAnchorGenerator/mul_8 (1/1 flops)
MultipleGridAnchorGenerator/mul_7 (1/1 flops)
MultipleGridAnchorGenerator/mul_6 (1/1 flops)
MultipleGridAnchorGenerator/mul_58 (1/1 flops)
MultipleGridAnchorGenerator/mul_57 (1/1 flops)
MultipleGridAnchorGenerator/mul_5 (1/1 flops)
MultipleGridAnchorGenerator/mul_4 (1/1 flops)
MultipleGridAnchorGenerator/mul_3 (1/1 flops)
MultipleGridAnchorGenerator/mul_2 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_18 (1/1 flops)
MultipleGridAnchorGenerator/mul_12 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_1 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_10 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_11 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_12 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_13 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_14 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_15 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_16 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_17 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/Greater_4 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_4 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_3 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_2 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_13 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_12 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_11 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_10 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_1 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/Greater_6 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/Greater_5 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_5 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/Greater_3 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/Greater_2 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/Greater_1 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/Greater (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_9 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_89 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_88 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_87 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_86 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_85 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_84 (1/1 flops)
MultipleGridAnchorGenerator/add_23 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_27 (1/1 flops)
MultipleGridAnchorGenerator/Minimum (1/1 flops)
Postprocessor/truediv_1 (1/1 flops)
Postprocessor/truediv (1/1 flops)
MultipleGridAnchorGenerator/add_15 (1/1 flops)
MultipleGridAnchorGenerator/add_16 (1/1 flops)
MultipleGridAnchorGenerator/add_18 (1/1 flops)
MultipleGridAnchorGenerator/add_19 (1/1 flops)
MultipleGridAnchorGenerator/add_20 (1/1 flops)
MultipleGridAnchorGenerator/add_21 (1/1 flops)
MultipleGridAnchorGenerator/add_22 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_83 (1/1 flops)
MultipleGridAnchorGenerator/assert_equal/Equal (1/1 flops)
MultipleGridAnchorGenerator/mul (1/1 flops)
MultipleGridAnchorGenerator/mul_1 (1/1 flops)
MultipleGridAnchorGenerator/mul_10 (1/1 flops)
MultipleGridAnchorGenerator/mul_11 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/mul (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_9 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_8 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_7 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/PadOrClipBoxList/sub_6 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_51 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_61 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_60 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_6 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_59 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_58 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_57 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_56 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_55 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_54 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_53 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_52 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_62 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_50 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_5 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_49 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_48 (1/1 flops)
MultipleGridAnchorGenerator/truediv_13 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_46 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_45 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_44 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_43 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_42 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_72 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_82 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_81 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_80 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_8 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_79 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_78 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_77 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_76 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_75 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_74 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_73 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_41 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_71 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_70 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_7 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_69 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_68 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_67 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_66 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_65 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_64 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/sub_63 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/SortByField_1/Equal (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_18 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_87 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_178 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_88 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_176 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_89 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_174 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_9 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_172 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_90 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_170 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/SortByField/Equal (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_86 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_168 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_166 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_26 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_164 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_10 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_162 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_100 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_160 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_16 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_25 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_80 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_4 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_76 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_38 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_77 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_36 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_78 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_34 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_79 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_32 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_8 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_30 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_158 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_81 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_28 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_82 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_26 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_83 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_24 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_84 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_22 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_85 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_20 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_2 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_13 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_19 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_134 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_132 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_1 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_130 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_10 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_11 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_128 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_12 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_126 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_136 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_124 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_14 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_122 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_15 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_120 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_12 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_16 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_118 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_17 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_116 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_18 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_108 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_102 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_156 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_24 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_154 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_104 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_152 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_23 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_150 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_106 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_22 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_148 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_40 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_146 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_21 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_144 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_20 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_142 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_110 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_140 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_14 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_2 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_138 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_112 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_48 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_39 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_4 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_40 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_41 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_42 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_43 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_44 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_45 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_46 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_98 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_47 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_96 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_38 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_94 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_49 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_92 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_5 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_90 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_50 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_51 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_88 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_52 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_86 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_53 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Greater (1/1 flops)
MultipleGridAnchorGenerator/truediv_11 (1/1 flops)
MultipleGridAnchorGenerator/truediv_12 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_114 (1/1 flops)
MultipleGridAnchorGenerator/truediv_2 (1/1 flops)
MultipleGridAnchorGenerator/truediv_3 (1/1 flops)
MultipleGridAnchorGenerator/truediv_4 (1/1 flops)
MultipleGridAnchorGenerator/truediv_5 (1/1 flops)
MultipleGridAnchorGenerator/truediv_6 (1/1 flops)
MultipleGridAnchorGenerator/truediv_7 (1/1 flops)
MultipleGridAnchorGenerator/truediv_8 (1/1 flops)
MultipleGridAnchorGenerator/truediv_9 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_84 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_28 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_29 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_3 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_30 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_31 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_32 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_33 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_34 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_35 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_36 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_37 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_7 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_65 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_60 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_6 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_66 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_58 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_67 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_56 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_68 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_54 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_69 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_52 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_62 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_50 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_70 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_71 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_48 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_72 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_46 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_73 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_44 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_74 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_42 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_75 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_59 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_54 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_82 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_55 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_80 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_8 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_56 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_78 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_57 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_76 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_58 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_74 (1/1 flops)
MultipleGridAnchorGenerator/truediv_10 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_72 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_6 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_70 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_60 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_61 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_68 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_62 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_66 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_63 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/add_64 (1/1 flops)
Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/Minimum_64 (1/1 flops)

======================End of Report==========================
W0521 17:26:52.311092 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:375: The name tf.train.Saver is deprecated. Please use tf.compat.v1.train.Saver instead.

2019-05-21 17:26:53.622503: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.623333: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:01:00.0
2019-05-21 17:26:53.623423: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.624338: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 1 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:02:00.0
2019-05-21 17:26:53.624346: I tensorflow/stream_executor/platform/default/dlopen_checker.cc:62] Not built with GPU enabled. Skip GPU library dlopen check.
2019-05-21 17:26:53.624389: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.625191: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.625994: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.626784: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.627569: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0, 1
2019-05-21 17:26:53.628033: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.628835: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:01:00.0
2019-05-21 17:26:53.628898: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.629711: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 1 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:02:00.0
2019-05-21 17:26:53.629718: I tensorflow/stream_executor/platform/default/dlopen_checker.cc:62] Not built with GPU enabled. Skip GPU library dlopen check.
2019-05-21 17:26:53.629775: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.630604: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.631462: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.632286: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.633101: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0, 1
2019-05-21 17:26:53.633165: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-05-21 17:26:53.633185: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187] 0 1
2019-05-21 17:26:53.633189: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0: N Y
2019-05-21 17:26:53.633208: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 1: Y N
2019-05-21 17:26:53.633741: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.634672: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.635454: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.636675: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9987 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
2019-05-21 17:26:53.636727: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:53.637950: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 10440 MB memory) -> physical GPU (device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:02:00.0, compute capability: 6.1)
W0521 17:26:53.638130 140549139449600 deprecation.py:323] From /home/zmsunnyday/anaconda3/envs/tf/lib/python3.6/site-packages/tensorflow/python/training/saver.py:1276: checkpoint_exists (from tensorflow.python.training.checkpoint_management) is deprecated and will be removed in a future version.
Instructions for updating:
Use standard file APIs to check for files with this prefix.
2019-05-21 17:26:56.698146: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.698937: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:01:00.0
2019-05-21 17:26:56.698995: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.699743: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 1 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:02:00.0
2019-05-21 17:26:56.699750: I tensorflow/stream_executor/platform/default/dlopen_checker.cc:62] Not built with GPU enabled. Skip GPU library dlopen check.
2019-05-21 17:26:56.699792: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.700543: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.701364: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.702188: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.702957: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0, 1
2019-05-21 17:26:56.703023: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-05-21 17:26:56.703042: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187] 0 1
2019-05-21 17:26:56.703046: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0: N Y
2019-05-21 17:26:56.703049: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 1: Y N
2019-05-21 17:26:56.703585: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.704352: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.705137: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.705883: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9987 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
2019-05-21 17:26:56.705949: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:56.706739: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 10440 MB memory) -> physical GPU (device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:02:00.0, compute capability: 6.1)
W0521 17:26:57.481124 140549139449600 deprecation.py:323] From /home/zmsunnyday/anaconda3/envs/tf/lib/python3.6/site-packages/tensorflow/python/tools/freeze_graph.py:233: convert_variables_to_constants (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.compat.v1.graph_util.convert_variables_to_constants
W0521 17:26:57.481256 140549139449600 deprecation.py:323] From /home/zmsunnyday/anaconda3/envs/tf/lib/python3.6/site-packages/tensorflow/python/framework/graph_util_impl.py:270: extract_sub_graph (from tensorflow.python.framework.graph_util_impl) is deprecated and will be removed in a future version.
Instructions for updating:
Use tf.compat.v1.graph_util.extract_sub_graph
2019-05-21 17:26:58.210823: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.211641: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:01:00.0
2019-05-21 17:26:58.211801: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.212577: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 1 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:02:00.0
2019-05-21 17:26:58.212585: I tensorflow/stream_executor/platform/default/dlopen_checker.cc:62] Not built with GPU enabled. Skip GPU library dlopen check.
2019-05-21 17:26:58.212659: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.213461: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.214266: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.215050: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.215828: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0, 1
2019-05-21 17:26:58.215894: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-05-21 17:26:58.215914: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187] 0 1
2019-05-21 17:26:58.215934: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0: N Y
2019-05-21 17:26:58.215937: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 1: Y N
2019-05-21 17:26:58.216441: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.217238: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.218040: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.218909: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9987 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
2019-05-21 17:26:58.218974: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:26:58.219781: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 10440 MB memory) -> physical GPU (device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:02:00.0, compute capability: 6.1)
W0521 17:26:59.623759 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:257: The name tf.saved_model.builder.SavedModelBuilder is deprecated. Please use tf.compat.v1.saved_model.builder.SavedModelBuilder instead.

W0521 17:26:59.624015 140549139449600 deprecation.py:323] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:260: build_tensor_info (from tensorflow.python.saved_model.utils_impl) is deprecated and will be removed in a future version.
Instructions for updating:
This function will only be available through the v1 compatibility library as tf.compat.v1.saved_model.utils.build_tensor_info or tf.compat.v1.saved_model.build_tensor_info.
W0521 17:26:59.624312 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:266: The name tf.saved_model.signature_def_utils.build_signature_def is deprecated. Please use tf.compat.v1.saved_model.signature_def_utils.build_signature_def instead.

W0521 17:26:59.624478 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/exporter.py:272: The name tf.saved_model.tag_constants.SERVING is deprecated. Please use tf.saved_model.SERVING instead.

W0521 17:27:00.251141 140549139449600 deprecation_wrapper.py:119] From /home/zmsunnyday/tensorrt/tftrt/examples/object_detection/third_party/models/research/object_detection/utils/config_util.py:180: The name tf.gfile.Open is deprecated. Please use tf.io.gfile.GFile instead.

2019-05-21 17:27:00.537818: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.538620: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:01:00.0
2019-05-21 17:27:00.538693: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.539544: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 1 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:02:00.0
2019-05-21 17:27:00.539551: I tensorflow/stream_executor/platform/default/dlopen_checker.cc:62] Not built with GPU enabled. Skip GPU library dlopen check.
2019-05-21 17:27:00.539607: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.540397: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.541214: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.542069: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.542833: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0, 1
2019-05-21 17:27:00.542898: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-05-21 17:27:00.542917: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187] 0 1
2019-05-21 17:27:00.542922: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0: N Y
2019-05-21 17:27:00.542925: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 1: Y N
2019-05-21 17:27:00.543479: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.544246: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.545050: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.545792: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9987 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
2019-05-21 17:27:00.545825: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:00.546590: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 10440 MB memory) -> physical GPU (device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:02:00.0, compute capability: 6.1)
W0521 17:27:00.546745 140549139449600 trt_convert.py:592] TensorRT mismatch. Compiled against version 5.0.2, but loaded 5.1.5. Things may not work
2019-05-21 17:27:02.677349: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.678657: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.679498: I tensorflow/core/grappler/devices.cc:55] Number of eligible GPUs (core count >= 8, compute capability >= 0.0): 2
2019-05-21 17:27:02.680142: I tensorflow/core/grappler/clusters/single_machine.cc:359] Starting new session
2019-05-21 17:27:02.681994: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.682754: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 0 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:01:00.0
2019-05-21 17:27:02.682800: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.683579: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1640] Found device 1 with properties:
name: GeForce GTX 1080 Ti major: 6 minor: 1 memoryClockRate(GHz): 1.582
pciBusID: 0000:02:00.0
2019-05-21 17:27:02.683586: I tensorflow/stream_executor/platform/default/dlopen_checker.cc:62] Not built with GPU enabled. Skip GPU library dlopen check.
2019-05-21 17:27:02.683642: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.684404: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.685231: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.685993: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.686762: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1763] Adding visible gpu devices: 0, 1
2019-05-21 17:27:02.686828: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1181] Device interconnect StreamExecutor with strength 1 edge matrix:
2019-05-21 17:27:02.686847: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1187] 0 1
2019-05-21 17:27:02.686851: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 0: N Y
2019-05-21 17:27:02.686870: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1200] 1: Y N
2019-05-21 17:27:02.687390: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.688175: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.689211: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.690096: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:0 with 9987 MB memory) -> physical GPU (device: 0, name: GeForce GTX 1080 Ti, pci bus id: 0000:01:00.0, compute capability: 6.1)
2019-05-21 17:27:02.690162: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:1005] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero
2019-05-21 17:27:02.691241: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1326] Created TensorFlow device (/job:localhost/replica:0/task:0/device:GPU:1 with 10440 MB memory) -> physical GPU (device: 1, name: GeForce GTX 1080 Ti, pci bus id: 0000:02:00.0, compute capability: 6.1)
2019-05-21 17:27:04.061111: I tensorflow/compiler/tf2tensorrt/segment/segment.cc:460] There are 1110 ops of 31 different types in the graph that are not converted to TensorRT: Identity, Slice, Range, GreaterEqual, Greater, Squeeze, Transpose, Pad, ConcatV2, ResizeBilinear, Reshape, Mul, Fill, Where, ExpandDims, NonMaxSuppressionV3, GatherV2, Unpack, Less, Cast, Minimum, Sum, Sub, Pack, TopKV2, Placeholder, Add, Shape, NoOp, StridedSlice, Select, (For more information see https://docs.nvidia.com/deeplearning/dgx/tf-trt-user-guide/index.html#supported-ops).
2019-05-21 17:27:04.371759: I tensorflow/compiler/tf2tensorrt/convert/convert_graph.cc:733] Number of TensorRT candidate segments: 2
2019-05-21 17:27:04.562027: I tensorflow/stream_executor/platform/default/dso_loader.cc:42] Successfully opened dynamic library libcudart.so.10.0
2019-05-21 17:27:04.562665: F tensorflow/compiler/tf2tensorrt/utils/trt_logger.cc:51] DefaultLoggerGot unknown severity level 4 from TensorRT: Plugin Creator registration succeeded - GridAnchor_TRT

why the run function parameter run_calibration was never used?

Hello, I am very curious that why the run_calibration parameter in the run method was never used?
So, I'm interested to see if the calibration function really works, especially if I switch to a different model

InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor

HI everyone. I try to optimize my frozen model, but i got below error. I dont know how to fix.
error:

InvalidArgumentError (see above for traceback): You must feed a value for placeholder tensor 'impor
t/inputs' with dtype float and shape [?,320,320,3]                                                 
         [[{{node import/inputs}} = Placeholder[dtype=DT_FLOAT, shape=[?,320,320,3], _device="/job:
localhost/replica:0/task:0/device:GPU:0"]()]]

My optimize code:

# Import TensorFlow and TensorRT                                           
import tensorflow as tf                                                    
import tensorflow.contrib.tensorrt as trt                                  
# Inference with TF-TRT frozen graph workflow:                             
graph = tf.Graph()                                                         
with graph.as_default():                                                   
    with tf.Session() as sess:                                             
        # First deserialize your frozen graph:                             
        with tf.gfile.GFile('frozen_darknet_yolov3_model.pb', 'rb') as f:  
            graph_def = tf.GraphDef()                                      
            graph_def.ParseFromString(f.read())                            
        # Now you can create a TensorRT inference graph from your          
        # frozen graph:                                                    
        trt_graph = trt.create_inference_graph(                            
            input_graph_def=graph_def,                                     
            outputs=['output_boxes'],                                      
            max_batch_size= 8 ,                                            
            max_workspace_size_bytes=24,                                   
            precision_mode="FP16")                                         
        # Import the TensorRT graph into a new graph and run:              
        output_node = tf.import_graph_def(                                 
            trt_graph,                                                     
            return_elements=["output_boxes"])                              
        sess.run(output_node)

how to convert a pre-trained model (frozen graph, savedmodel or ckpt) to a new tf-trt-optimized savedmodel for tensorflow:serving

System information

OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Linux RHEL 7.6
TensorFlow installed from (source or binary): binary
TensorFlow version (use command below):1.13.1
Python version: Python 2.7.16 :: Anaconda, Inc.
CUDA/cuDNN version: CUDA 10.1, cuDNN 7.6
GPU model and memory: 2 Tesla K80, 12G/12G

Describe the current behavior

Use the tool saved_model_cli to covert a severable model with TF-TRT:

saved_model_cli convert \
--dir "/path/to/mask_rcnn/saved_model" \
--output_dir "/path/to/trt-mask-rcnn" \
--tag_set serve \
tensorrt --precision_mode FP32 --max_batch_size 32 --is_dynamic_op True

While I got a .pb graph and the variables fold is empty, which can't be launched on tensorflow:serving.

variables/
    NULL
saved_model.pb

Describe the expected behavior

The following structure is expected. https://www.tensorflow.org/guide/saved_model#structure_of_a_savedmodel_directory

variables/
    variables.data-?????-of-?????
    variables.index
saved_model.pb|saved_model.pbtxt

My final target is to Optimizing TensorFlow Serving performance with NVIDIA TensorRT, and the model that I want to launch is a pre-trained mask-rcnn downloaded from tensorflow_model_zoo

I have successfully launched several models without TensorRT, while how can I accelerate the inference with the help of TensorRT.

Any help would be grateful!

release time

hi, when will community finalize tf-trt and release a stable version?

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.

Jobs

Jooble