add centernet dla34 ctdet task. (wang-xinyu#558)

* add centernet dla34 ctdet task.

* update readme.

* update readme and fix a bug in sample.

Co-authored-by: chandler <[email protected]>

Author: SsisyphusTao

centernet/README.md

@@ -0,0 +1,29 @@
+# CenterNet
+
+This is the trt implementation of detection model [ctdet_coco_dla_2x](https://drive.google.com/open?id=1pl_-ael8wERdUREEnaIfqOV_VF2bEVRT) from [xingyizhou/CenterNet](https://github.com/xingyizhou/CenterNet) official work. 
+
+## How to Run
+
+1. Follow [NVIDIA/TensorRT](https://github.com/NVIDIA/TensorRT) tutorial to build TensorRT7
+
+2. Copy folder `dcnv2Plugin` to `TensorRT/plugin` and edit `InferPlugin.cpp` and `CMakeLists.txt`
+
+3. Rebuild to install custom plugin
+
+4. Use `tensorrt-7.2.3.4-cp36-none-linux_x86_64.whl` in TensorRT OSS to update your python-tensorrt
+
+5. Run `python centernet.py -m ${PTH_PATH} -s` to create trt engine 
+
+## Sample
+
+```
+// Download ctdet_coco_dla_2x.pth and transfer it into trt engine first
+// Download the test img from https://raw.githubusercontent.com/tensorflow/models/master/research/deeplab/g3doc/img/image2.jpg or choose your own one
+cd sample
+python test.py ${ENGINE_PATH} ${IMG_PATH}
+```
+![trt_out](https://user-images.githubusercontent.com/47047345/119128637-7a878900-ba68-11eb-91ff-5dcc10f01b77.jpg)
+
+## TODO
+
+Integrate the post process with trt engine to make it more easier to use.

centernet/centernet.py

@@ -0,0 +1,21 @@
+#
+# Copyright (c) 2020, NVIDIA CORPORATION. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+#
+file(GLOB SRCS *.cpp)
+set(PLUGIN_SOURCES ${PLUGIN_SOURCES} ${SRCS})
+set(PLUGIN_SOURCES ${PLUGIN_SOURCES} PARENT_SCOPE)
+file(GLOB CU_SRCS *.cu)
+set(PLUGIN_CU_SOURCES ${PLUGIN_CU_SOURCES} ${CU_SRCS})
+set(PLUGIN_CU_SOURCES ${PLUGIN_CU_SOURCES} PARENT_SCOPE)

centernet/dcnv2Plugin/CMakeLists.txt

@@ -0,0 +1,100 @@
+/*!
+ ******************* BEGIN Caffe Copyright Notice and Disclaimer ****************
+ *
+ * COPYRIGHT
+ *
+ * All contributions by the University of California:
+ * Copyright (c) 2014-2017 The Regents of the University of California (Regents)
+ * All rights reserved.
+ *
+ * All other contributions:
+ * Copyright (c) 2014-2017, the respective contributors
+ * All rights reserved.
+ *
+ * Caffe uses a shared copyright model: each contributor holds copyright over
+ * their contributions to Caffe. The project versioning records all such
+ * contribution and copyright details. If a contributor wants to further mark
+ * their specific copyright on a particular contribution, they should indicate
+ * their copyright solely in the commit message of the change when it is
+ * committed.
+ *
+ * LICENSE
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions are met:
+ *
+ * 1. Redistributions of source code must retain the above copyright notice, this
+ * list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright notice,
+ * this list of conditions and the following disclaimer in the documentation
+ * and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
+ * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
+ * DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE LIABLE FOR
+ * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+ * SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * CONTRIBUTION AGREEMENT
+ *
+ * By contributing to the BVLC/caffe repository through pull-request, comment,
+ * or otherwise, the contributor releases their content to the
+ * license and copyright terms herein.
+ *
+ ***************** END Caffe Copyright Notice and Disclaimer ********************
+ *
+ * Copyright (c) 2018 Microsoft
+ * Licensed under The MIT License [see LICENSE for details]
+ * \file modulated_deformable_im2col.h
+ * \brief Function definitions of converting an image to
+ * column matrix based on kernel, padding, dilation, and offset.
+ * These functions are mainly used in deformable convolution operators.
+ * \ref: https://arxiv.org/abs/1811.11168
+ * \author Yuwen Xiong, Haozhi Qi, Jifeng Dai, Xizhou Zhu, Han Hu
+ */
+
+/***************** Adapted by Charles Shang *********************/
+
+#ifndef DCN_V2_IM2COL_CUDA
+#define DCN_V2_IM2COL_CUDA
+
+// #ifdef __cplusplus
+// extern "C"
+// {
+// #endif
+
+  void modulated_deformable_im2col_cuda(cudaStream_t stream,
+                                        const float *data_im, const float *data_offset, const float *data_mask,
+                                        const int batch_size, const int channels, const int height_im, const int width_im,
+                                        const int height_col, const int width_col, const int kernel_h, const int kenerl_w,
+                                        const int pad_h, const int pad_w, const int stride_h, const int stride_w,
+                                        const int dilation_h, const int dilation_w,
+                                        const int deformable_group, float *data_col);
+
+  void modulated_deformable_col2im_cuda(cudaStream_t stream,
+                                        const float *data_col, const float *data_offset, const float *data_mask,
+                                        const int batch_size, const int channels, const int height_im, const int width_im,
+                                        const int height_col, const int width_col, const int kernel_h, const int kenerl_w,
+                                        const int pad_h, const int pad_w, const int stride_h, const int stride_w,
+                                        const int dilation_h, const int dilation_w,
+                                        const int deformable_group, float *grad_im);
+
+  void modulated_deformable_col2im_coord_cuda(cudaStream_t stream,
+                                         const float *data_col, const float *data_im, const float *data_offset, const float *data_mask,
+                                         const int batch_size, const int channels, const int height_im, const int width_im,
+                                         const int height_col, const int width_col, const int kernel_h, const int kenerl_w,
+                                         const int pad_h, const int pad_w, const int stride_h, const int stride_w,
+                                         const int dilation_h, const int dilation_w,
+                                         const int deformable_group,
+                                         float *grad_offset, float *grad_mask);
+
+// #ifdef __cplusplus
+// }
+// #endif
+
+#endif