forked from shouxieai/tensorRT_Pro
-
Notifications
You must be signed in to change notification settings - Fork 1
/
Makefile
200 lines (153 loc) · 5.59 KB
/
Makefile
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
cpp_srcs := $(shell find src -name "*.cpp")
cpp_objs := $(cpp_srcs:.cpp=.o)
cpp_objs := $(cpp_objs:src/%=objs/%)
cpp_mk := $(cpp_objs:.o=.mk)
cu_srcs := $(shell find src -name "*.cu")
cu_objs := $(cu_srcs:.cu=.cuo)
cu_objs := $(cu_objs:src/%=objs/%)
cu_mk := $(cu_objs:.cuo=.cumk)
# 配置你的库路径
# 1. onnx-tensorrt（项目集成了，不需要配置，下面地址是下载位置）
# https://github.com/onnx/onnx-tensorrt/tree/release/8.0
# 2. protobuf（请自行下载编译）
# https://github.com/protocolbuffers/protobuf/tree/v3.11.4
# 3. cudnn8.2.2.26（请自行下载）
# runtime的tar包，runtime中包含了lib、so文件
# develop的tar包，develop中包含了include、h等文件
# 4. tensorRT-8.0.1.6-cuda10.2（请自行下载）
# tensorRT下载GA版本（通用版、稳定版），EA（尝鲜版本）不要
# 5. cuda10.2，也可以是11.x看搭配（请自行下载安装）
lean_protobuf := /data/sxai/lean/protobuf3.11.4
lean_tensor_rt := /data/sxai/lean/TensorRT-8.0.1.6-cuda10.2-cudnn8.2
lean_cudnn := /data/sxai/lean/cudnn8.2.2.26
lean_opencv := /data/sxai/lean/opencv4.2.0
lean_cuda := /data/sxai/lean/cuda-10.2
use_python := true
python_root := /data/datav/newbb/lean/anaconda3/envs/torch1.8
python_name := python3.9
include_paths := src \
src/application \
src/tensorRT \
src/tensorRT/common \
$(lean_protobuf)/include \
$(lean_opencv)/include/opencv4 \
$(lean_tensor_rt)/include \
$(lean_cuda)/include \
$(lean_cudnn)/include
library_paths := $(lean_protobuf)/lib \
$(lean_opencv)/lib \
$(lean_tensor_rt)/lib \
$(lean_cuda)/lib64 \
$(lean_cudnn)/lib \
/datav/k12/lean/ffmpeg4.2/lib
link_librarys := opencv_core opencv_imgproc opencv_videoio opencv_imgcodecs \
nvinfer nvinfer_plugin \
cuda cublas cudart cudnn \
stdc++ protobuf dl
# HAS_PYTHON表示是否编译python支持
support_define :=
ifeq ($(use_python), true)
include_paths += $(python_root)/include/$(python_name)
library_paths += $(python_root)/lib
link_librarys += $(python_name)
support_define += -DHAS_PYTHON
endif
paths := $(foreach item,$(library_paths),-Wl,-rpath=$(item))
include_paths := $(foreach item,$(include_paths),-I$(item))
library_paths := $(foreach item,$(library_paths),-L$(item))
link_librarys := $(foreach item,$(link_librarys),-l$(item))
# 如果是其他显卡，请修改-gencode=arch=compute_75,code=sm_75为对应显卡的能力
# 显卡对应的号码参考这里：https://developer.nvidia.com/zh-cn/cuda-gpus#compute
# 如果是jetson nano，提示找不到-m64指令，请删掉 -m64选项。不影响结果
cpp_compile_flags := -std=c++11 -fPIC -m64 -g -fopenmp -w -O0 $(support_define)
cu_compile_flags := -std=c++11 -m64 -Xcompiler -fPIC -g -w -gencode=arch=compute_75,code=sm_75 -O0 $(support_define)
link_flags := -pthread -fopenmp -Wl,-rpath='$$ORIGIN'
cpp_compile_flags += $(include_paths)
cu_compile_flags += $(include_paths)
link_flags += $(library_paths) $(link_librarys) $(paths)
ifneq ($(MAKECMDGOALS), clean)
-include $(cpp_mk) $(cu_mk)
endif
pro : workspace/pro
trtpyc : python/trtpy/libtrtpyc.so
workspace/pro : $(cpp_objs) $(cu_objs)
@echo Link $@
@mkdir -p $(dir $@)
@g++ $^ -o $@ $(link_flags)
python/trtpy/libtrtpyc.so : $(cpp_objs) $(cu_objs)
@echo Link $@
@mkdir -p $(dir $@)
@g++ -shared $^ -o $@ $(link_flags)
objs/%.o : src/%.cpp
@echo Compile CXX $<
@mkdir -p $(dir $@)
@g++ -c $< -o $@ $(cpp_compile_flags)
objs/%.cuo : src/%.cu
@echo Compile CUDA $<
@mkdir -p $(dir $@)
@nvcc -c $< -o $@ $(cu_compile_flags)
objs/%.mk : src/%.cpp
@echo Compile depends CXX $<
@mkdir -p $(dir $@)
@g++ -M $< -MF $@ -MT $(@:.mk=.o) $(cpp_compile_flags)
objs/%.cumk : src/%.cu
@echo Compile depends CUDA $<
@mkdir -p $(dir $@)
@nvcc -M $< -MF $@ -MT $(@:.cumk=.o) $(cu_compile_flags)
yolo : workspace/pro
@cd workspace && ./pro yolo
yolo_fast : workspace/pro
@cd workspace && ./pro yolo_fast
bert : workspace/pro
@cd workspace && ./pro bert
alphapose : workspace/pro
@cd workspace && ./pro alphapose
fall : workspace/pro
@cd workspace && ./pro fall_recognize
retinaface : workspace/pro
@cd workspace && ./pro retinaface
arcface : workspace/pro
@cd workspace && ./pro arcface
arcface_video : workspace/pro
@cd workspace && ./pro arcface_video
arcface_tracker : workspace/pro
@cd workspace && ./pro arcface_tracker
test_all : workspace/pro
@cd workspace && ./pro test_all
scrfd : workspace/pro
@cd workspace && ./pro scrfd
centernet : workspace/pro
@cd workspace && ./pro centernet
dbface : workspace/pro
@cd workspace && ./pro dbface
high_perf : workspace/pro
@cd workspace && ./pro high_perf
lesson : workspace/pro
@cd workspace && ./pro lesson
plugin : workspace/pro
@cd workspace && ./pro plugin
pytorch : trtpyc
@cd python && python test_torch.py
pyscrfd : trtpyc
@cd python && python test_scrfd.py
pyretinaface : trtpyc
@cd python && python test_retinaface.py
pycenternet : trtpyc
@cd python && python test_centernet.py
pyyolov5 : trtpyc
@cd python && python test_yolov5.py
pyyolox : trtpyc
@cd python && python test_yolox.py
pyinstall : trtpyc
@cd python && python setup.py install
debug :
@echo $(includes)
clean :
@rm -rf objs workspace/pro python/trtpy/libtrtpyc.so python/build python/dist python/trtpy.egg-info python/trtpy/__pycache__
@rm -rf workspace/single_inference
@rm -rf workspace/scrfd_result workspace/retinaface_result
@rm -rf workspace/YoloV5_result workspace/YoloX_result
@rm -rf workspace/face/library_draw workspace/face/result
@rm -rf build
@rm -rf python/trtpy/libplugin_list.so
.PHONY : clean yolo alphapose fall debug