forked from Tianxiaomo/pytorch-YOLOv4
-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_inference.py
199 lines (181 loc) · 8.47 KB
/
test_inference.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
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
import torch
import argparse
import os
import random
import shutil
import time
import warnings
from torch import nn
import torch.nn.functional as F
from tool.torch_utils import *
from tool.yolo_layer import YoloLayer
from yolo_loss import Yolo_loss
from torch.utils import mkldnn as mkldnn_utils
from models import Yolov4
from dataset import Yolo_dataset
parser = argparse.ArgumentParser(description='PyTorch Yolov4 Training')
parser.add_argument('-N', '--n_classes', default=80, type=int, metavar='n_classes',
help='num classes')
parser.add_argument('-w', '--weightfile', type=str, default='./yolov4.pth',
help='weight file')
parser.add_argument('-i', '--imgfile', type=str, default='./data/dog.jpg',
help='image file')
parser.add_argument('--height', default=320, type=int, help='height')
parser.add_argument('--width', default=320, type=int, help='width')
parser.add_argument('-n', '--namesfile', type=str, help='names file')
parser.add_argument('-j', '--workers', default=4, type=int, metavar='N',
help='number of data loading workers (default: 4)')
parser.add_argument('--ipex', action='store_true', default=False,
help='use intel pytorch extension')
parser.add_argument('--int8', action='store_true', default=False,
help='enable ipex int8 path')
parser.add_argument('--bf16', action='store_true', default=False,
help='enable ipex bf16 path')
parser.add_argument('--jit', action='store_true', default=False,
help='enable ipex jit fusionpath')
parser.add_argument('--calibration', action='store_true', default=False,
help='doing calibration step for int8 path')
parser.add_argument('--configure-dir', default='configure.json', type=str, metavar='PATH',
help = 'path to int8 configures, default file name is configure.json')
parser.add_argument("--dummy", action='store_true',
help="using dummu data to test the performance of inference")
parser.add_argument('--warmup', default=30, type=int, metavar='N',
help='number of warmup iterati ons to run')
parser.add_argument('-b', '--batch-size', default=64, type=int, metavar='N',
help='mini-batch size (default: 64), this is the total '
'batch size of all GPUs on the current node when '
'using Data Parallel or Distributed Data Parallel')
def fp32_imperative_stock_pytorch(model, images):
model = model.to(memory_format=torch.channels_last)
images = images.to(memory_format=torch.channels_last)
output=model(images)
print(output)
def fp32_jit_stock_pytorch(model, images, args):
model = torch.jit.trace(model, torch.randn(args.batch_size, 3, args.height, args.width))
model = torch.jit.freeze(model)
images = images.to(memory_format=torch.channels_last)
output=model(images)
print(output)
def bf16_imperative_stock_pytorch(model, images):
model = model.to(torch.bfloat16).to(memory_format=torch.channels_last)
with torch.cpu.amp.autocast():
images = images.to(torch.bfloat16).to(memory_format=torch.channels_last)
output=model(images)
print(output)
def bf16_jit_stock_pytorch(model, images, args):
model = model.to(torch.bfloat16).to(memory_format=torch.channels_last)
with torch.no_grad():
with torch.cpu.amp.autocast():
model = torch.jit.trace(model, torch.randn(args.batch_size, 3, args.height, args.width))
model = torch.jit.freeze(model)
with torch.cpu.amp.autocast():
images = images.to(torch.bfloat16).to(memory_format=torch.channels_last)
output=model(images)
print(output)
def fp32_imperative_stock_pytorch_ipex(model, images):
import intel_pytorch_extension as ipex
model = model.to(memory_format=torch.channels_last)
model = ipex.optimize(model, dtype=torch.float32, level="O0")
images = images.to(memory_format=torch.channels_last)
output=model(images)
print(output)
def fp32_jit_stock_pytorch_ipex(model, images, args):
import intel_pytorch_extension as ipex
model = model.to(memory_format=torch.channels_last)
model = ipex.optimize(model, dtype=torch.float32, level="O0")
model = torch.jit.trace(model, torch.rand(args.batch_size, 3, args.height, args.width).to(memory_format=torch.channels_last))
model = torch.jit.freeze(model)
images = images.contiguous(memory_format=torch.channels_last)
output=model(images)
print(output)
def bf16_imperative_stock_pytorch_ipex(model, images):
import intel_pytorch_extension as ipex
model = model.to(torch.bfloat16).to(memory_format=torch.channels_last)
model = ipex.optimize(model, dtype=torch.bfloat16, level="O0")
conf = ipex.AmpConf(torch.bfloat16)
with ipex.amp.autocast(enabled=True, configure=conf), torch.no_grad():
output=model(images)
print(output)
def bf16_jit_stock_pytorch_ipex(model, images, args):
import intel_pytorch_extension as ipex
with torch.no_grad():
model = model.to(torch.bfloat16).to(memory_format=torch.channels_last)
model = ipex.optimize(model, dtype=torch.bfloat16, level="O0")
conf = ipex.AmpConf(torch.bfloat16)
images = images.to(torch.bfloat16)
with ipex.amp.autocast(enabled=True, configure=conf), torch.no_grad():
model = torch.jit.trace(model, torch.rand(args.batch_size, 3, args.height, args.width).to(memory_format=torch.channels_last))
model = torch.jit.freeze(model)
output=model(images)
print(output)
def fp32_imperative_pytorch_ipex(model, images):
import intel_pytorch_extension as ipex
model = model.to(ipex.DEVICE)
images = images.to(ipex.DEVICE)
output=model(images)
print(output)
def fp32_jit_pytorch_ipex(model, images, args):
import intel_pytorch_extension as ipex
model = model.to(ipex.DEVICE)
images = images.to(ipex.DEVICE)
with torch.no_grad():
model = torch.jit.trace(model, torch.randn(args.batch_size, 3, args.height, args.width).to(ipex.DEVICE))
output=model(images)
print(output)
def bf16_imperative_pytorch_ipex(model, images):
import intel_pytorch_extension as ipex
ipex.enable_auto_mixed_precision(mixed_dtype = torch.bfloat16)
model = model.to(ipex.DEVICE)
images = images.to(ipex.DEVICE)
output=model(images)
print(output)
def bf16_jit_pytorch_ipex(model, images, args):
import intel_pytorch_extension as ipex
ipex.enable_auto_mixed_precision(mixed_dtype = torch.bfloat16)
model = model.to(ipex.DEVICE)
images = images.to(ipex.DEVICE)
with torch.no_grad():
model = torch.jit.trace(model, torch.randn(args.batch_size, 3, args.height, args.width).to(ipex.DEVICE))
output=model(images)
print(output)
def get_data(args):
if args.dummy:
images = torch.randn(args.batch_size, 3, args.height, args.width)
return images
else:
import sys
import cv2
img = cv2.imread(args.imgfile)
sized = cv2.resize(img, (args.width, args.height))
img = cv2.cvtColor(sized, cv2.COLOR_BGR2RGB)
if type(img) == np.ndarray and len(img.shape) == 3: # cv2 image
img = torch.from_numpy(img.transpose(2, 0, 1)).float().div(255.0).unsqueeze(0)
elif type(img) == np.ndarray and len(img.shape) == 4:
img = torch.from_numpy(img.transpose(0, 3, 1, 2)).float().div(255.0)
else:
print("unknow image type")
exit(-1)
img = torch.autograd.Variable(img)
return img
if __name__ == "__main__":
args = parser.parse_args()
print(args)
n_classes = args.n_classes
weightfile = args.weightfile
model = Yolov4(yolov4conv137weight=None, n_classes=n_classes, inference=True)
pretrained_dict = torch.load(weightfile, map_location=torch.device('cpu'))
model.load_state_dict(pretrained_dict)
img = get_data(args)
model.eval()
fp32_imperative_stock_pytorch(model, img)
fp32_jit_stock_pytorch(model, img, args)
bf16_imperative_stock_pytorch(model, img)
bf16_jit_stock_pytorch(model, img, args)
fp32_imperative_stock_pytorch_ipex(model, img)
fp32_jit_stock_pytorch_ipex(model, img, args)
bf16_imperative_stock_pytorch_ipex(model, img)
bf16_jit_stock_pytorch_ipex(model, img, args)
fp32_imperative_pytorch_ipex(model, img)
fp32_jit_pytorch_ipex(model, img, args)
bf16_imperative_pytorch_ipex(model, img)
bf16_jit_pytorch_ipex(model, img, args)