thatbrguy/pedestrian-detection

Pedestrian detection using the TensorFlow Object Detection API. Includes multi GPU parallel processing inference.

Pedestrian-Detection

Pedestrian Detection using the TensorFlow Object Detection API and Nanonets. [Blog][Performance]

This repo provides complementary material to this blog post, which compares the performance of four object detectors for a pedestrian detection task. It also introduces a feature to use multiple GPUs in parallel for inference using the multiprocessing package. The count accuracy and FPS for different models (using 1,2,4 or 8 GPUs in parallel) were calculated and plotted.

Dataset

The TownCentre dataset is used for training our pedestrian detector. You can use the following commands to download the dataset. This automatically extracts the frames from the video, and creates XML files from the csv groundtruth. The image dimensions are downscaled by a factor of 2 to reduce processing overhead.

wget http://www.robots.ox.ac.uk/ActiveVision/Research/Projects/2009bbenfold_headpose/Datasets/TownCentreXVID.avi
wget http://www.robots.ox.ac.uk/ActiveVision/Research/Projects/2009bbenfold_headpose/Datasets/TownCentre-groundtruth.top
python extract_towncentre.py
python extract_GT.py

Setup

1. For TensorFlow Object Detection API

Refer to the instructions in this blog post.

2. For Nanonets

Step 1: Clone the repo

git clone https://github.com/NanoNets/object-detection-sample-python.git
cd object-detection-sample-python
sudo pip install requests

Step 2: Get your free API Key

Get your free API Key from http://app.nanonets.com/user/api_key

Step 3: Set the API key as an Environment Variable

export NANONETS_API_KEY=YOUR_API_KEY_GOES_HERE

Step 4: Create a New Model

python ./code/create-model.py

Note: An environment variable NANONETS_MODEL_ID will be created in the previous step, with your model ID.

Step 5: Upload the Training Data

Place the training data in a folder named images and annotations in annotations/json

python ./code/upload-training.py

Step 6: Train the Model

python ./code/train-model.py

Step 7: Get Model State

The model takes ~2 hours to train. You will get an email once the model is trained. In the meanwhile you check the state of the model

python ./code/model-state.py

Step 8: Make Predictions

Create a folder named 'test_images' inside the 'nanonets' folder. Place the input images in this folder, and then run this command.

python ./code/prediction.py

Results

FPS vs GPUs

For more stats, refer to the blog post. The performance of each model (on the test set) was compiled into a video, which you can see here.

In light of GDPR and feeble accountability of Deep Learning, it is imperative that we ponder about the legality and ethical issues concerning automation of surveillance. This blog/code is for educational purposes only, and it used a publicly available dataset. It is your responsibility to make sure that your automated system complies with the law in your region.

pedestrian-detection's Issues

when i run inference.py, i got this issues

Traceback (most recent call last):
File "C:\ProgramData\Anaconda3\envs\tensorflow\lib\multiprocessing\process.py", line 258, in _bootstrap
self.run()
File "C:\ProgramData\Anaconda3\envs\tensorflow\lib\multiprocessing\process.py", line 93, in run
self._target(*self._args, **self._kwargs)
File "E:\Pedestrian-Detector-master\object_detection\inference.py", line 55, in feed
TEST_IMAGE_PATHS, RESULT_IMAGE_PATHS, _ = load_details(args)
NameError: name 'args' is not defined

this is my code:
python object_detection/inference.py
--input_dir=test_images
--output_dir=test_result
--label_map=annotations/label_map.pbtxt
--frozen_graph=object_detection/pedestrian_pb/frozen_inference_graph.pb
--num_output_classes=1
--n_jobs=1
--delay=0
i do not knwo how to fix it , could you help me

Is there a limit to the number of bounding boxes ?

Hello,

When I extracted the bounding boxes of a crowded scene, I found that the array associated to the bounding boxes has got a limit of 100 rows. Does it mean that if an image has more than 100 pedestrians the detector won't put them all in bounding boxes ?

How about training ? Can we train a model with a dataset of images that contains more than 100 pedestrians ?

Thank you in advance for your clarifications !

error when run python create_tf_record.py --data_dir=`pwd` --output_dir=`pwd`

Hi,
When I run python create_tf_record.py --data_dir=pwd --output_dir=pwd command, I found error:
File "create_tf_record.py", line 180, in <module> tf.app.run() File "/home/administrator/.local/lib/python2.7/site-packages/tensorflow/python/platform/app.py", line 125, in run _sys.exit(main(argv)) File "create_tf_record.py", line 175, in main image_dir, train_examples) File "create_tf_record.py", line 147, in create_tf_record tf_example = dict_to_tf_example(data, label_map_dict, image_dir) File "create_tf_record.py", line 60, in dict_to_tf_example img_path = os.path.join(image_subdirectory, data['filename']) TypeError: string indices must be integers, not str

The problem of protobuf

I run the train.py by python3 , and meet the below error. I try transfer the string into bytes, but it no work. so have anyone have any advice.

Traceback (most recent call last):
File "object_detection/train.py", line 53, in
from builders import model_builder
File "/wangpdm/Pedestrian-Detection/object_detection/builders/model_builder.py", line 17, in
from builders import anchor_generator_builder
File "/wangpdm/Pedestrian-Detection/object_detection/builders/anchor_generator_builder.py", line 20, in
from protos import anchor_generator_pb2
File "/wangpdm/Pedestrian-Detection/object_detection/protos/anchor_generator_pb2.py", line 16, in
from protos import grid_anchor_generator_pb2 as object__detection_dot_protos_dot_grid__anchor__generator__pb2
File "/wangpdm/Pedestrian-Detection/object_detection/protos/grid_anchor_generator_pb2.py", line 22, in
serialized_pb=_b('\n3object_detection/protos/grid_anchor_generator.proto\x12\x17 protos"\xcd\x01\n\x13GridAnchorGenerator\x12\x13\n\x06height\x18\x01 \x01(\x05:\x03\x32\x35\x36\x12\x12\n\x05width\x18\x02 \x01(\x05:\x03\x32\x35\x36\x12\x19\n\rheight_stride\x18\x03 \x01(\x05:\x02\x31\x36\x12\x18\n\x0cwidth_stride\x18\x04 \x01(\x05:\x02\x31\x36\x12\x18\n\rheight_offset\x18\x05 \x01(\x05:\x01\x30\x12\x17\n\x0cwidth_offset\x18\x06 \x01(\x05:\x01\x30\x12\x0e\n\x06scales\x18\x07 \x03(\x02\x12\x15\n\raspect_ratios\x18\x08 \x03(\x02')
File "/usr/local/lib/python3.6/dist-packages/google/protobuf/descriptor.py", line 879, in new
return _message.default_pool.AddSerializedFile(serialized_pb)
TypeError: Couldn't parse file content!

error

Hello, when I wsas using your model, I found that there was an error in creating the model, the response was [500], do you now what caused that? Thanks.

How to run the code on any videos?

Hi, This is awesome!
Can you please help to figure it out how I have to use this code to detect pedestrians on any videos? I am very new to Python!

Thank you so much!
Kasra,

Training on Colab, how to pass arguments to the .ipynb on colab

Hello,

Following your advices, I am going to train a model on Colab.
However, I am currently facing a small problem, I don't know how to pass arguments to my .ipynb.
Can you tell me how you did it ?

By the way, do I need to convert all the .py files called by the train script to .ipynb or is it enough to have only the main file with .ipynb extension (train.py, in this case) ?

Thank you for your help !

Which parameters to reduce to avoid ResourceExhaustedError

Hello, I try to train the faster_rcnn_inception_resnet_v2_atrous_coco_2018_01_28 model by launching the train.py script on it, but I get the following ResourceExhaustedError. Do you have any idea on how to solve it ? I tried to change many parameters in pipeline.config, but It doesn't change anything

2018-10-10 14:54:05.313837: I tensorflow/core/common_runtime/bfc_allocator.cc:678] Sum Total of in-use chunks: 1.25GiB
2018-10-10 14:54:05.313845: I tensorflow/core/common_runtime/bfc_allocator.cc:680] Stats:
Limit: 1363345408
InUse: 1338755072
MaxInUse: 1350130944
NumAllocs: 3937
MaxAllocSize: 256131072

2018-10-10 14:54:05.313921: W tensorflow/core/common_runtime/bfc_allocator.cc:279] ****************************************************************************************************
2018-10-10 14:54:05.313944: W tensorflow/core/framework/op_kernel.cc:1275] OP_REQUIRES failed at transpose_op.cc:199 : Resource exhausted: OOM when allocating tensor with shape[4,160,42,64] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
INFO:tensorflow:Error reported to Coordinator: OOM when allocating tensor with shape[4,160,42,64] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[Node: FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/Conv2D-0-TransposeNHWCToNCHW-LayoutOptimizer = Transpose[T=DT_FLOAT, Tperm=DT_INT32, _device="/job:localhost/replica:0/task:0/device:GPU:0"](FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/SpaceToBatchND, PermConstNHWCToNCHW-LayoutOptimizer)]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

[[Node: BatchMultiClassNonMaxSuppression/map/while/MultiClassNonMaxSuppression/strided_slice/_1871 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_11146...ided_slice", tensor_type=DT_INT32, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.
Traceback (most recent call last):
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1278, in _do_call
return fn(*args)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1263, in _run_fn
options, feed_dict, fetch_list, target_list, run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1350, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.ResourceExhaustedError: OOM when allocating tensor with shape[4,160,42,64] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[Node: FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/Conv2D-0-TransposeNHWCToNCHW-LayoutOptimizer = Transpose[T=DT_FLOAT, Tperm=DT_INT32, _device="/job:localhost/replica:0/task:0/device:GPU:0"](FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/SpaceToBatchND, PermConstNHWCToNCHW-LayoutOptimizer)]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

[[Node: BatchMultiClassNonMaxSuppression/map/while/MultiClassNonMaxSuppression/strided_slice/_1871 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_11146...ided_slice", tensor_type=DT_INT32, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/coordinator.py", line 297, in stop_on_exception
yield
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/coordinator.py", line 495, in run
self.run_loop()
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/supervisor.py", line 1035, in run_loop
self._sv.global_step])
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 877, in run
run_metadata_ptr)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1100, in _run
feed_dict_tensor, options, run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1272, in _do_run
run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1291, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.ResourceExhaustedError: OOM when allocating tensor with shape[4,160,42,64] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[Node: FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/Conv2D-0-TransposeNHWCToNCHW-LayoutOptimizer = Transpose[T=DT_FLOAT, Tperm=DT_INT32, _device="/job:localhost/replica:0/task:0/device:GPU:0"](FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/SpaceToBatchND, PermConstNHWCToNCHW-LayoutOptimizer)]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

[[Node: BatchMultiClassNonMaxSuppression/map/while/MultiClassNonMaxSuppression/strided_slice/_1871 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_11146...ided_slice", tensor_type=DT_INT32, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

Traceback (most recent call last):
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1278, in _do_call
return fn(*args)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1263, in _run_fn
options, feed_dict, fetch_list, target_list, run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1350, in _call_tf_sessionrun
run_metadata)
tensorflow.python.framework.errors_impl.ResourceExhaustedError: OOM when allocating tensor with shape[1,384,72,128] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[Node: FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_1a_3x3/Conv2D = Conv2D[T=DT_FLOAT, data_format="NCHW", dilations=[1, 1, 1, 1], padding="SAME", strides=[1, 1, 1, 1], use_cudnn_on_gpu=true, _device="/job:localhost/replica:0/task:0/device:GPU:0"](FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_0b_3x3/Relu, FirstStageFeatureExtractor/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_1a_3x3/weights/read/_3137)]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

[[Node: gradients/SecondStageFeatureExtractor/InceptionResnetV2/Repeat/block8_9/Conv2d_1x1/Conv2D_grad/tuple/control_dependency_1/_5073 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_13509...pendency_1", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/supervisor.py", line 995, in managed_session
yield sess
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/contrib/slim/python/slim/learning.py", line 770, in train
sess, train_op, global_step, train_step_kwargs)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/contrib/slim/python/slim/learning.py", line 487, in train_step
run_metadata=run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 877, in run
run_metadata_ptr)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1100, in _run
feed_dict_tensor, options, run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1272, in _do_run
run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1291, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.ResourceExhaustedError: OOM when allocating tensor with shape[1,384,72,128] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[Node: FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_1a_3x3/Conv2D = Conv2D[T=DT_FLOAT, data_format="NCHW", dilations=[1, 1, 1, 1], padding="SAME", strides=[1, 1, 1, 1], use_cudnn_on_gpu=true, _device="/job:localhost/replica:0/task:0/device:GPU:0"](FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_0b_3x3/Relu, FirstStageFeatureExtractor/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_1a_3x3/weights/read/_3137)]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

[[Node: gradients/SecondStageFeatureExtractor/InceptionResnetV2/Repeat/block8_9/Conv2d_1x1/Conv2D_grad/tuple/control_dependency_1/_5073 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_13509...pendency_1", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

Caused by op 'FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_1a_3x3/Conv2D', defined at:
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/train.py", line 163, in
tf.app.run()
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 125, in run
_sys.exit(main(argv))
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/train.py", line 159, in main
worker_job_name, is_chief, FLAGS.train_dir)
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/trainer.py", line 228, in train
clones = model_deploy.create_clones(deploy_config, model_fn, [input_queue])
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/deployment/model_deploy.py", line 193, in create_clones
outputs = model_fn(*args, **kwargs)
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/trainer.py", line 165, in _create_losses
prediction_dict = detection_model.predict(images)
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/meta_architectures/faster_rcnn_meta_arch.py", line 531, in predict
image_shape) = self._extract_rpn_feature_maps(preprocessed_inputs)
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/meta_architectures/faster_rcnn_meta_arch.py", line 685, in _extract_rpn_feature_maps
preprocessed_inputs, scope=self.first_stage_feature_extractor_scope)
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/meta_architectures/faster_rcnn_meta_arch.py", line 134, in extract_proposal_features
return self._extract_proposal_features(preprocessed_inputs, scope)
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/models/faster_rcnn_inception_resnet_v2_feature_extractor.py", line 112, in _extract_proposal_features
align_feature_maps=True))
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/nets/inception_resnet_v2.py", line 232, in inception_resnet_v2_base
scope='Conv2d_1a_3x3')
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/contrib/framework/python/ops/arg_scope.py", line 183, in func_with_args
return func(*args, **current_args)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/contrib/layers/python/layers/layers.py", line 1154, in convolution2d
conv_dims=2)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/contrib/framework/python/ops/arg_scope.py", line 183, in func_with_args
return func(*args, **current_args)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/contrib/layers/python/layers/layers.py", line 1057, in convolution
outputs = layer.apply(inputs)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/keras/engine/base_layer.py", line 805, in apply
return self.call(inputs, *args, **kwargs)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/layers/base.py", line 362, in call
outputs = super(Layer, self).call(inputs, *args, **kwargs)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/keras/engine/base_layer.py", line 736, in call
outputs = self.call(inputs, *args, **kwargs)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/keras/layers/convolutional.py", line 186, in call
outputs = self._convolution_op(inputs, self.kernel)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py", line 868, in call
return self.conv_op(inp, filter)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py", line 520, in call
return self.call(inp, filter)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/ops/nn_ops.py", line 204, in call
name=self.name)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/ops/gen_nn_ops.py", line 956, in conv2d
data_format=data_format, dilations=dilations, name=name)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/framework/op_def_library.py", line 787, in _apply_op_helper
op_def=op_def)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/util/deprecation.py", line 454, in new_func
return func(*args, **kwargs)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 3155, in create_op
op_def=op_def)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/framework/ops.py", line 1717, in init
self._traceback = tf_stack.extract_stack()

ResourceExhaustedError (see above for traceback): OOM when allocating tensor with shape[1,384,72,128] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[Node: FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_1a_3x3/Conv2D = Conv2D[T=DT_FLOAT, data_format="NCHW", dilations=[1, 1, 1, 1], padding="SAME", strides=[1, 1, 1, 1], use_cudnn_on_gpu=true, _device="/job:localhost/replica:0/task:0/device:GPU:0"](FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_0b_3x3/Relu, FirstStageFeatureExtractor/InceptionResnetV2/Mixed_6a/Branch_1/Conv2d_1a_3x3/weights/read/_3137)]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

[[Node: gradients/SecondStageFeatureExtractor/InceptionResnetV2/Repeat/block8_9/Conv2d_1x1/Conv2D_grad/tuple/control_dependency_1/_5073 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_13509...pendency_1", tensor_type=DT_FLOAT, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/train.py", line 163, in
tf.app.run()
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/platform/app.py", line 125, in run
_sys.exit(main(argv))
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/train.py", line 159, in main
worker_job_name, is_chief, FLAGS.train_dir)
File "/home/mounir/PycharmProjects/Pedestrian-Detection-master/object_detection/trainer.py", line 332, in train
saver=saver)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/contrib/slim/python/slim/learning.py", line 785, in train
ignore_live_threads=ignore_live_threads)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/contextlib.py", line 99, in exit
self.gen.throw(type, value, traceback)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/supervisor.py", line 1005, in managed_session
self.stop(close_summary_writer=close_summary_writer)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/supervisor.py", line 833, in stop
ignore_live_threads=ignore_live_threads)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/coordinator.py", line 389, in join
six.reraise(*self._exc_info_to_raise)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/six.py", line 693, in reraise
raise value
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/coordinator.py", line 297, in stop_on_exception
yield
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/coordinator.py", line 495, in run
self.run_loop()
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/training/supervisor.py", line 1035, in run_loop
self._sv.global_step])
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 877, in run
run_metadata_ptr)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1100, in _run
feed_dict_tensor, options, run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1272, in _do_run
run_metadata)
File "/home/mounir/anaconda3/envs/tflow-gpu/lib/python3.6/site-packages/tensorflow/python/client/session.py", line 1291, in _do_call
raise type(e)(node_def, op, message)
tensorflow.python.framework.errors_impl.ResourceExhaustedError: OOM when allocating tensor with shape[4,160,42,64] and type float on /job:localhost/replica:0/task:0/device:GPU:0 by allocator GPU_0_bfc
[[Node: FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/Conv2D-0-TransposeNHWCToNCHW-LayoutOptimizer = Transpose[T=DT_FLOAT, Tperm=DT_INT32, _device="/job:localhost/replica:0/task:0/device:GPU:0"](FirstStageFeatureExtractor/InceptionResnetV2/InceptionResnetV2/Repeat_1/block17_2/Branch_1/Conv2d_0c_7x1/SpaceToBatchND, PermConstNHWCToNCHW-LayoutOptimizer)]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

[[Node: BatchMultiClassNonMaxSuppression/map/while/MultiClassNonMaxSuppression/strided_slice/_1871 = _Recvclient_terminated=false, recv_device="/job:localhost/replica:0/task:0/device:CPU:0", send_device="/job:localhost/replica:0/task:0/device:GPU:0", send_device_incarnation=1, tensor_name="edge_11146...ided_slice", tensor_type=DT_INT32, _device="/job:localhost/replica:0/task:0/device:CPU:0"]]
Hint: If you want to see a list of allocated tensors when OOM happens, add report_tensor_allocations_upon_oom to RunOptions for current allocation info.

Process finished with exit code 1

Performance issue in the definition of testResizeToRangePreservesStaticSpatialShape, object_detection/core/preprocessor_test.py(P1)

Hello, I found a performance issue in the definition of testResizeToRangePreservesStaticSpatialShape, object_detection/core/preprocessor_test.py, tf.shape(out_image) will created repeatedly during program execution, resulting in reduced efficiency. I think in_image, out_image and out_image_shape should be created before the loop in testResizeToRangeWithDynamicSpatialShape.

Looking forward to your reply. Btw, I am very glad to create a PR to fix it if you are too busy.

Hello, is your pre-train model easy to provide?

Full pipeline.config

Would you have available the full pipeline.config file? It looks like it has been clipped in the blog post.