hosang / gossipnet Goto Github PK

View Code? Open in Web Editor NEW

79.0 4.0 25.0 89 KB

Non-maximum suppression for object detection in a neural network

License: Other

Python 71.40% Makefile 0.49% C++ 21.57% Cuda 5.69% C 0.86%

tensorflow object-detection non-maximum-suppression graph-convolutional-networks

gossipnet's People

Contributors

Stargazers

Watchers

gossipnet's Issues

pretrained model

Hi @hosang,

Where is the pretrained model define in .yaml pretrained_model: /BS/bbox_nms_net2/work/tf-models/resnet_v1_101.ckpt

Thanks

Trained Model for COCO multiclass setting

Hi @hosang,

Thanks for providing the code! Will it be possible for you to provide the trained model for COCO multiclass setting, I am trying to incorporate image features also in the network and having the trained model will help me compare the results.

Thanks!

Run with TF 1.12.0

I successfully run the code with TF 1.12.0
You need run two sh instead of the make file.
In roi_pooling_layer, please run the sh file

TF_INC=$(python -c 'import tensorflow as tf; print(tf.sysconfig.get_include())')

TF_LIB=$(python -c 'import tensorflow as tf; print(tf.sysconfig.get_lib())')

echo $TF_INC

echo $TF_LIB

CUDA_PATH=/usr/local/cuda/

g++ -std=c++11 -c roi_pooling_op.cc -o roi_pooling_op.o -fPIC -I "/root/anaconda2/envs/ld_tf_base/lib/python3.6/site-packages/tensorflow/include" -O2

nvcc -std=c++11 -c -o roi_pooling_op.cu.o roi_pooling_op_gpu.cu -I $TF_INC -D GOOGLE_CUDA=1 -x cu -Xcompiler -fPIC -arch=sm_52

## if you install tf using already-built binary, or gcc version 4.x, uncomment the two lines below

#g++ -std=c++11 -shared -D_GLIBCXX_USE_CXX11_ABI=0 -o roi_pooling.so roi_pooling_op.cc \

# roi_pooling_op.cu.o -I $TF_INC -fPIC -lcudart -L $CUDA_PATH/lib64

# for gcc5-built tf

g++ -std=c++11 -shared -D_GLIBCXX_USE_CXX11_ABI=0 -o roi_pooling.so roi_pooling_op.cc roi_pooling_op.cu.o -I $TF_INC -I $TF_INC"/external/nsync/public" -L $TF_LIB -ltensorflow_framework -O2 -fPIC -lcudart -L $CUDA_PATH/lib64

cd ..

in matching_module run the sh file

TF_INC=$(python -c 'import tensorflow as tf; print(tf.sysconfig.get_include())')

TF_LIB=$(python -c 'import tensorflow as tf; print(tf.sysconfig.get_lib())')

echo $TF_INC

echo $TF_LIB

CUDA_PATH=/usr/local/cuda/

g++ -std=c++11 -c det_matching.cc -o det_matching.o -fPIC -I "/root/anaconda2/envs/ld_tf_base/lib/python3.6/site-packages/tensorflow/include" -O2

# nvcc -std=c++11 -c -o det_matching.o det_matching.cc -I $TF_INC -D GOOGLE_CUDA=1 -x cu -Xcompiler -fPIC -arch=sm_52

## if you install tf using already-built binary, or gcc version 4.x, uncomment the two lines below

#g++ -std=c++11 -shared -D_GLIBCXX_USE_CXX11_ABI=0 -o roi_pooling.so roi_pooling_op.cc \

# roi_pooling_op.cu.o -I $TF_INC -fPIC -lcudart -L $CUDA_PATH/lib64

# for gcc5-built tf

g++ -std=c++11 -shared -D_GLIBCXX_USE_CXX11_ABI=0 -o det_matching.so det_matching.cc -I $TF_INC -I $TF_INC"/external/nsync/public" -L $TF_LIB -ltensorflow_framework -O2 -fPIC -lcudart -L $CUDA_PATH/lib64

cd ..

download the coco json annotation in the direction ProjectDir/data/coco/annotations

instances_minival2014.json
instances_train2014.json
instances_val2014.json
instances_valminusminival2014.json

Please train

python train.py --config=experiments/coco_person/conf.yaml

the result like this:

2020-08-17 03:18:11.800835  iter  59940   lr   0.0001   opt loss  1.90954     data loss normalized 0.0163604   unnormalized  1.76164
2020-08-17 03:18:11.940283  iter  59960   lr   0.0001   opt loss  6.11693     data loss normalized 0.032737   unnormalized  5.96905
2020-08-17 03:18:12.069166  iter  59980   lr   0.0001   opt loss 0.713293     data loss normalized 0.00841691   unnormalized 0.565408
2020-08-17 03:18:12.189150  iter  60000   lr   0.0001   opt loss  3.09831     data loss normalized 0.0365992   unnormalized  2.95046
2020-08-17 03:18:12.189251  starting validation
[1]
2020-08-17 03:18:27.687428  iter  60000   validation pass:   mAP  66.7   multiclass AP  66.7
/root/ld/PycharmProjects/gossipnet/outputs/gnet
wrote model to /root/ld/PycharmProjects/gossipnet/outputs/gnet-60000
Iteration   mAP   
     20000    64.0
     40000    65.9
     60000    66.7  (best)

roi_pooling.so: undefined symbol

Hi @hosang,

I got the following error:

tensorflow.python.framework.errors_impl.NotFoundError: /work/DL/code/gossipnet/nms_net/roi_pooling_layer/roi_pooling.so: undefined symbol: _ZN10tensorflow7strings6StrCatB5cxx11ERKNS0_8AlphaNumE

roi_pooling.so Makefile

Hello, i get the following error: tensorflow.python.framework.errors_impl.NotFoundError: ~/nms_net/roi_pooling_layer/roi_pooling.so: undefined symbol: _ZN10tensorflow7strings6StrCatB5cxx11ERKNS0_8AlphaNumE Is there any advice solving it?

Problem with Makefile

Hello!
I did make and in terminal I just saw the warnings, no errors. But when I tried to debug train.py I get such error:
tensorflow.python.framework.errors_impl.NotFoundError: ~/nms_net/roi_pooling_layer/roi_pooling.so: undefined symbol: _ZN10tensorflow7strings6StrCatB5cxx11ERKNS0_8AlphaNumE

Thanks!

AttributeError: 'EasyDict' object has no attribute 'test_model'

Python 2.7.6
tf.version '1.2.1'
make passed.
After changed some code has been renamed in tensorflow, such as tf.pack to tf.stack, tf.select to tf.where and so on.
I got this error:
restorer.restore(sess, cfg.test_model)
AttributeError: 'EasyDict' object has no attribute 'test_model'
run log as:

python test.py -c ./experiments/coco_person/conf.yaml OUTPUT
reading /mnt/geekvc/gossipnet/data/cache/coco_2014_minival_FRCN_train_imdb_cache.pkl
preparing test imdb
5000 images: 1821412 detections, 446 crowd annotations, 35821 non-crowd annotations
dropping all classes but person
5000 images: 337940 detections, 227 crowd annotations, 10633 non-crowd annotations
dropping images without detections
4429 images: 337940 detections, 227 crowd annotations, 10628 non-crowd annotations
done
doing single class NMS
WARNING:tensorflow:From /mnt/geekvc/gossipnet/nms_net/network.py:314: add_loss (from tensorflow.contrib.framework.python.ops.arg_scope) is deprecated and will be removed after 2016-12-30.
Instructions for updating:
Use tf.losses.add_loss instead.
2017-10-31 11:12:05.534766: I tensorflow/core/common_runtime/gpu/gpu_device.cc:955] Found device 0 with properties:
name: Tesla K40c
major: 3 minor: 5 memoryClockRate (GHz) 0.8755
pciBusID 0000:02:00.0
Total memory: 12.00GiB
Free memory: 7.67GiB
2017-10-31 11:12:05.696399: W tensorflow/stream_executor/cuda/cuda_driver.cc:523] A non-primary context 0x74929e0 exists before initializing the StreamExecutor. We haven't verified StreamExecutor works with that.
2017-10-31 11:12:05.697048: I tensorflow/core/common_runtime/gpu/gpu_device.cc:955] Found device 1 with properties:
name: Tesla K40c
major: 3 minor: 5 memoryClockRate (GHz) 0.8755
pciBusID 0000:03:00.0
Total memory: 12.00GiB
Free memory: 11.72GiB
2017-10-31 11:12:05.880469: W tensorflow/stream_executor/cuda/cuda_driver.cc:523] A non-primary context 0x6ef5220 exists before initializing the StreamExecutor. We haven't verified StreamExecutor works with that.
2017-10-31 11:12:05.881353: I tensorflow/core/common_runtime/gpu/gpu_device.cc:955] Found device 2 with properties:
name: Tesla K40c
major: 3 minor: 5 memoryClockRate (GHz) 0.8755
pciBusID 0000:83:00.0
Total memory: 12.00GiB
Free memory: 11.72GiB
2017-10-31 11:12:06.069418: W tensorflow/stream_executor/cuda/cuda_driver.cc:523] A non-primary context 0x6189f80 exists before initializing the StreamExecutor. We haven't verified StreamExecutor works with that.
2017-10-31 11:12:06.070325: I tensorflow/core/common_runtime/gpu/gpu_device.cc:955] Found device 3 with properties:
name: Tesla K40c
major: 3 minor: 5 memoryClockRate (GHz) 0.8755
pciBusID 0000:84:00.0
Total memory: 12.00GiB
Free memory: 11.72GiB
2017-10-31 11:12:06.070608: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 0 and 2
2017-10-31 11:12:06.070671: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 0 and 3
2017-10-31 11:12:06.070776: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 1 and 2
2017-10-31 11:12:06.070835: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 1 and 3
2017-10-31 11:12:06.070892: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 2 and 0
2017-10-31 11:12:06.070948: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 2 and 1
2017-10-31 11:12:06.071131: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 3 and 0
2017-10-31 11:12:06.071190: I tensorflow/core/common_runtime/gpu/gpu_device.cc:847] Peer access not supported between device ordinals 3 and 1
2017-10-31 11:12:06.071425: I tensorflow/core/common_runtime/gpu/gpu_device.cc:976] DMA: 0 1 2 3
2017-10-31 11:12:06.071465: I tensorflow/core/common_runtime/gpu/gpu_device.cc:986] 0:   Y Y N N
2017-10-31 11:12:06.071498: I tensorflow/core/common_runtime/gpu/gpu_device.cc:986] 1:   Y Y N N
2017-10-31 11:12:06.071534: I tensorflow/core/common_runtime/gpu/gpu_device.cc:986] 2:   N N Y Y
2017-10-31 11:12:06.071568: I tensorflow/core/common_runtime/gpu/gpu_device.cc:986] 3:   N N Y Y
2017-10-31 11:12:06.071615: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1045] Creating TensorFlow device (/gpu:0) -> (device: 0, name: Tesla K40c, pci bus id: 0000:02:00.0)
2017-10-31 11:12:06.071656: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1045] Creating TensorFlow device (/gpu:1) -> (device: 1, name: Tesla K40c, pci bus id: 0000:03:00.0)
2017-10-31 11:12:06.071691: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1045] Creating TensorFlow device (/gpu:2) -> (device: 2, name: Tesla K40c, pci bus id: 0000:83:00.0)
2017-10-31 11:12:06.071729: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1045] Creating TensorFlow device (/gpu:3) -> (device: 3, name: Tesla K40c, pci bus id: 0000:84:00.0)
Traceback (most recent call last):
  File "test.py", line 135, in <module>
    main()
  File "test.py", line 131, in main
    dets = test_run(test_imdb)
  File "test.py", line 61, in test_run
    restorer.restore(sess, cfg.test_model)
AttributeError: 'EasyDict' object has no attribute 'test_model'

Thank you very in advance!

do you have caffe version ?

KeyError: 'gt_classes'

Hi @hosang,

I got following error:
File "train.py", line 87, in load_and_enqueue
food = {ph: batch[name] for (name, ph) in placeholders}
File "train.py", line 87, in
food = {ph: batch[name] for (name, ph) in placeholders}
KeyError: 'gt_classes'

look like batch has key name gt_classes

Strong Tnet

I'm interested in your article，At the same time, I am very interested in the "A CONVNET FOR NON-MAXIMUM SUPPRESSION ", Can you share “Strong Tnet“” code for me？
@hosang

there is an error running train.py

Traceback (most recent call last):
File "train.py", line 17, in
import imdb
File "/home/zhangyb/DL/Detection/gossipnet/imdb/init.py", line 13, in
from imdb.pal import load_pal
File "/home/zhangyb/DL/Detection/gossipnet/imdb/pal.py", line 11, in
from imdb.file_formats import pal
File "/home/zhangyb/DL/Detection/gossipnet/imdb/file_formats/pal.py", line 5, in
from . import AnnoList_pb2
ImportError: cannot import name AnnoList_pb2

do you have any suggestions? thanks in advance....

compute Map

Hi,

I am investigating the code, but it seems that Map calculation is weird.

it seems that you did not use iou thresh as defined here https://github.com/hosang/gossipnet/blob/master/nms_net/config.py#L63 .

Am I right?

Error in compiling: make: *** [nms_net/matching_module/det_matching.so] Fehler 1

Hi @hosang,

I tried to compile the repo, but I am getting the following error.
(objdetTF3) azim_se@othrys2:~/azim_se/projects/detection/gossipnet$ make
g++ -std=c++11 -shared nms_net/matching_module/det_matching.cc -o nms_net/matching_module/det_matching.so -fPIC -I "/home/azim_se/.virtualenvs/objdetTF3/lib/python3.5/site-packages/tensorflow/include" -O2
In file included from nms_net/matching_module/det_matching.cc:5:0:
nms_net/matching_module/det_matching.cc: In member function ‘void DetectionMatchingOp::Compute(tensorflow::OpKernelContext*)’:
/home/azim_se/.virtualenvs/objdetTF3/lib/python3.5/site-packages/tensorflow/include/tensorflow/core/framework/op_kernel.h:1538:3: error: expected ‘;’ before ‘do’
do {
^
nms_net/matching_module/det_matching.cc:91:5: note: in expansion of macro ‘OP_REQUIRES’
OP_REQUIRES(context, ious.dimension(1) == ignore.dimension(0),
^
nms_net/matching_module/det_matching.cc:95:5: error: expected ‘;’ before ‘argsort’
argsort(score, &det_order);
^
Makefile:19: die Regel für Ziel „nms_net/matching_module/det_matching.so“ scheiterte
make: *** [nms_net/matching_module/det_matching.so] Fehler 1

Here is my system configuration:
Ubuntu 16.04
Tensorflow 1.8
Cuda 9.0
Cudnn 7.0
GPU: Titan X Pascal
Python 3.5.2

Here is the system path configuration:
echo $PATH
/home/azim_se/.virtualenvs/objdetTF3/bin:/usr/local/cuda-9.0/bin:/home/azim_se/MyProjects/pytorch/eigen3/compiled:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:/snap/bin:/usr/lib/jvm/java-8-oracle/bin:/usr/lib/jvm/java-8-oracle/db/bin:/usr/lib/jvm/java-8-oracle/jre/bin:/home/azim_se/softwares/pycharm/bin:/home/azim_se/softwares/eclipse:/home/azim_se/MyProjects/bin:/home/azim_se/MATLAB/R2016b/bin:/home/azim_se/.local/bin:/home/azim_se/softwares/tensorRT/4.0.0.3-ubuntu16.04-cuda9.0-cudnn7.0.5/bin
echo $LD_LIBRARY_PATH
/home/azim_se/softwares/cuDNN/9.0-v7.0.5-ubuntu16.04/lib64:/home/azim_se/MyProjects/pytorch/eigen3:/home/azim_se/MyProjects/pytorch/eigen3/build:/usr/local/cuda-9.0/lib64:/usr/local/opencv-3.4.1/lib/x86_64-linux-gnu:/home/azim_se/softwares/tensorRT/4.0.0.3-ubuntu16.04-cuda9.0-cudnn7.0.5/lib:/home/azim_se/.local/lib:/home/azim_se/MyProjects/pytorch/nccl/compiled/lib
echo $PYTHONPATH
/home/azim_se/projects/detection/ModelsTF/research:/home/azim_se/projects/detection//ModelsTF/research/slim

hosang / gossipnet Goto Github PK

gossipnet's People

Contributors

Stargazers

Watchers

Forkers

gossipnet's Issues

Recommend Projects

Recommend Topics

Recommend Org