forked from DataXujing/TensorRT-DETR
-
Notifications
You must be signed in to change notification settings - Fork 1
/
performance_accuracy_detr.py
190 lines (122 loc) · 5.52 KB
/
performance_accuracy_detr.py
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
#
# Copyright (c) 2021, 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.
#
# ~~~Medcare AI Lab~~~
import os
import torch
import torchvision
from torchsummary import summary
import time
import pycuda.driver as cuda
import pycuda.autoinit
import numpy as np
from trt_util.process_img import PyTorchTensorHolder
from trt_util.trt_lite import TrtLite
from trt_util.process_img import preprocess_torch_v1,preprocess_torch_v2
import matplotlib.pyplot as plt
import matplotlib
matplotlib.use('Agg')
from model.hubconf import detr_resnet50
device = torch.device("cuda:0")
# model torch
model_path='checkpoint/detr_resnet50.pth'
detr = detr_resnet50(pretrained=False,num_classes=20+1).eval() # <------这里类别需要+1
state_dict = torch.load(model_path) # <-----------修改加载模型的路径
detr.load_state_dict(state_dict["model"])
detr.to(device)
# model trt fp32
engine_path="./detr.plan"
trt_32 = TrtLite(engine_file_path=engine_path)
# model trt fp16
engine_path_16="./detr_fp16.plan"
trt_16 = TrtLite(engine_file_path=engine_path_16)
def acc_static_torch(test_dir,detr=detr):
img, _, _ = preprocess_torch_v2(test_dir)
outputs = detr(img.to(device))
scores = outputs['pred_logits'].cpu().detach().numpy()
boxes = outputs['pred_boxes'].cpu().detach().numpy()
return scores.ravel(), boxes.ravel()
def acc_static_trt(test_dir,trt_=trt_32):
# trt_.print_info()
i2shape = {0: (1, 3, 800, 800)}
io_info = trt_.get_io_info(i2shape)
d_buffers = trt_.allocate_io_buffers(i2shape,True)
output_data_trt_prob = np.zeros(io_info[1][2], dtype=np.float32)
output_data_trt_box = np.zeros(io_info[2][2], dtype=np.float32)
img, _, _ = preprocess_torch_v2(test_dir)
d_buffers[0] = img.to(device)
trt_.execute([t.data_ptr() for t in d_buffers], i2shape)
scores = d_buffers[1].cpu().numpy().ravel()
boxes = d_buffers[2].cpu().numpy().ravel()
return scores, boxes
if __name__ == "__main__":
# 平均相对精度的计算
# np.mean(np.abs(output_data_pytorch - output_data_trt) / np.abs(output_data_pytorch))
files = os.listdir("./test")
Average_diff_perc_socre_32 = []
Average_diff_perc_box_32 = []
Average_diff_perc_socre_16 = []
Average_diff_perc_box_16 = []
for file in files[:1000]:
print(file)
file_path = os.path.join("./test/",file)
torch_score, torch_box = acc_static_torch(file_path)
# fp32
trt_score_32,trt_box_32 = acc_static_trt(file_path,trt_=trt_32)
# fp16
trt_score_16,trt_box_16 = acc_static_trt(file_path,trt_=trt_16)
adp_score_32 = np.mean(np.abs(torch_score - trt_score_32) / np.abs(torch_score))
adp_box_32 = np.mean(np.abs(torch_box - trt_box_32) / np.abs(torch_box))
adp_score_16 = np.mean(np.abs(torch_score - trt_score_16) / np.abs(torch_score))
adp_box_16 = np.mean(np.abs(torch_box - trt_box_16) / np.abs(torch_box))
Average_diff_perc_socre_32.append(adp_score_32)
Average_diff_perc_box_32.append(adp_box_32)
Average_diff_perc_socre_16.append(adp_score_16)
Average_diff_perc_box_16.append(adp_box_16)
print("-"*50)
print(f"trt FP32 Score的平均相对精度：{Average_diff_perc_socre_32}")
print(f"trt FP32 Box的平均相对精度：{Average_diff_perc_box_32}")
print(f"trt FP16 Score的平均相对精度：{Average_diff_perc_socre_16}")
print(f"trt FP16 Box的平均相对精度：{Average_diff_perc_box_16}")
print("-"*50)
# plot Average diff percentage
plt.rcParams['figure.figsize'] = (16.0, 9.0) # 单位是inches
fig,subs=plt.subplots(2,2)
subs[0][0].plot(np.arange(len(files[:1000])), Average_diff_perc_socre_32, 'ro',label='FP32(Score)')
subs[0][0].axhline(y=1e-6,color='b',linestyle='--')
subs[0][0].axhline(y=1e-5,color='b',linestyle='--')
subs[0][0].set_xlabel('Test Image ID')
subs[0][0].set_ylabel('Average Diff Percentage')
subs[0][0].legend()
subs[0][1].plot(np.arange(len(files[:1000])), Average_diff_perc_box_32, 'bv',label='FP32(Box)')
subs[0][1].axhline(y=1e-6,color='r',linestyle='--')
subs[0][1].axhline(y=1e-5,color='r',linestyle='--')
subs[0][1].set_xlabel('Test Image ID')
subs[0][1].set_ylabel('Average Diff Percentage')
subs[0][1].legend()
subs[1][0].plot(np.arange(len(files[:1000])), Average_diff_perc_socre_16, 'g^',label='FP16(Score)')
subs[1][0].axhline(y=1e-3,color='k',linestyle='--')
subs[1][0].axhline(y=1e-2,color='k',linestyle='--')
subs[1][0].set_xlabel('Test Image ID')
subs[1][0].set_ylabel('Average Diff Percentage')
subs[1][0].legend()
subs[1][1].plot(np.arange(len(files[:1000])), Average_diff_perc_box_16, 'k*',label='FP16(Box)')
subs[1][1].axhline(y=1e-3,color='g',linestyle='--')
subs[1][1].axhline(y=1e-2,color='g',linestyle='--')
subs[1][1].set_xlabel('Test Image ID')
subs[1][1].set_ylabel('Average Diff Percentage')
subs[1][1].legend()
plt.savefig("./average_diff_percentage.png")
plt.close()