-
Notifications
You must be signed in to change notification settings - Fork 8
/
quantize.py
211 lines (183 loc) · 8.21 KB
/
quantize.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
199
200
201
202
203
204
205
206
207
208
209
210
211
import argparse
from utee import misc, quant, selector
import torch
import torch.backends.cudnn as cudnn
cudnn.benchmark =True
from collections import OrderedDict
import pprint
import os
known_models = [
'mnist', 'svhn', # 28x28
'cifar10', 'cifar100', # 32x32
'stl10', # 96x96
'alexnet', # 224x224
'vgg16', 'vgg16_bn', 'vgg19', 'vgg19_bn', # 224x224
'resnet18', 'resnet34', 'resnet50', 'resnet101','resnet152', # 224x224
'squeezenet_v0', 'squeezenet_v1', #224x224
'inception_v3', # 299x299
]
parser = argparse.ArgumentParser(description='PyTorch Quantization')
parser.add_argument('--use_model_zoo', type=int, default=1, help='decide if use model_zoo')
parser.add_argument('--type', default='MobileNetV2', help='|'.join(selector.known_models))
parser.add_argument('--data_root', default='~/dataset', help='folder to save the model')
parser.add_argument('--model_root', default='~/pytorch-mobilenet-v2/mobilenetv2_718.pth', help='the path of pre-trained parammeters')
parser.add_argument('--net_root', default='~/pytorch-mobilenet-v2/MobileNetV2.py', help='the path of pre-trained parammeters')
parser.add_argument('--test', type=int, default=1, help='test data distribution')
parser.add_argument('--batch_size', type=int, default=1, help='input batch size for training')
parser.add_argument('--n_sample', type=int, default=10, help='number of samples to infer the scaling factor')
parser.add_argument('--gpu', default="0", help='index of gpus to use')
parser.add_argument('--ngpu', type=int, default=1, help='number of gpus to use')
parser.add_argument('--logdir', default='log/default', help='folder to save to the log')
parser.add_argument('--replace_bn', type=int, default=0, help='decide if replace bn layer')
parser.add_argument('--map_bn', type=int, default=1, help='decide if map bn layer to conv layer')
parser.add_argument('--input_size', type=int, default=224, help='input size of image')
parser.add_argument('--shuffle', type=int, default=1, help='data shuffle')
parser.add_argument('--overflow_rate', type=float, default=0.0, help='overflow rate')
parser.add_argument('--quant_method', default='linear', help='linear|minmax|log|tanh|scale')
parser.add_argument('--param_bits', type=int, default=8, help='bit-width for parameters')
parser.add_argument('--bn_bits', type=int, default=8, help='bit-width for running mean and std')
parser.add_argument('--fwd_bits', type=int, default=8, help='bit-width for layer output')
args = parser.parse_args()
args.gpu = misc.auto_select_gpu(utility_bound=0, num_gpu=args.ngpu, selected_gpus=args.gpu)
args.ngpu = len(args.gpu)
misc.ensure_dir(args.logdir)
args.model_root = misc.expand_user(args.model_root)
args.data_root = misc.expand_user(args.data_root)
args.input_size = 299 if 'inception' in args.type else args.input_size
assert args.quant_method in ['linear', 'minmax', 'log', 'tanh','scale']
print("=================FLAGS==================")
for k, v in args.__dict__.items():
print('{}: {}'.format(k, v))
print("========================================")
assert torch.cuda.is_available(), 'no cuda'
#torch.manual_seed(args.seed)
#torch.cuda.manual_seed(args.seed)
# load model and dataset fetcher
if args.use_model_zoo:
args.model_root = os.path.expanduser('~/.torch/models/')
model, ds_fetcher, is_imagenet = selector.select(model_name=args.type, model_root=args.model_root)
args.ngpu = args.ngpu if is_imagenet else 1
else:
args.model_root = '~/pytorch-mobilenet-v2/mobilenetv2_718.pth'
model, ds_fetcher = selector.find(
model_name = args.type,
model_root = args.model_root,
net_root = args.net_root)
# replace bn with 1x1 conv
if args.replace_bn:
quant.replace_bn(model)
# map bn to conv
if args.map_bn:
quant.bn2conv(model)
# quantize parameters
print("=================quantize parameters==================")
if args.param_bits < 32:
state_dict = model.state_dict()
state_dict_quant = OrderedDict()
sf_dict = OrderedDict()
for k, v in state_dict.items():
if 'running' in k: # quantize bn layer
#print("k:{}, v:\n{}".format(k,v))
if args.bn_bits >=32:
print("Ignoring {}".format(k))
state_dict_quant[k] = v
continue
else:
bits = args.bn_bits
else:
bits = args.param_bits
if args.quant_method == 'linear':
sf = bits - 1. - quant.compute_integral_part(v, overflow_rate=args.overflow_rate)
# sf stands for float bits
v_quant = quant.linear_quantize(v, sf, bits=bits)
#if 'bias' in k:
#print("{}, sf:{}, quantized value:\n{}".format(k,sf, v_quant.sort(dim=0, descending=True)[0]))
elif args.quant_method == 'log':
v_quant = quant.log_minmax_quantize(v, bits=bits)
elif args.quant_method == 'minmax':
v_quant = quant.min_max_quantize(v, bits=bits)
else:
v_quant = quant.tanh_quantize(v, bits=bits)
state_dict_quant[k] = v_quant
print("k={0:<35}, bits={1:<5}, sf={2:d>}".format(k,bits,sf))
model.load_state_dict(state_dict_quant)
print("======================================================")
# quantize forward activation
print("=================quantize activation==================")
if args.fwd_bits < 32:
model = quant.duplicate_model_with_quant(model,
bits=args.fwd_bits,
overflow_rate=args.overflow_rate,
counter=args.n_sample,
type=args.quant_method)
# ds_fetcher is in path: /imagenet/dataset.get
val_ds_tmp = ds_fetcher(batch_size=args.batch_size,
data_root=args.data_root,
train=False,
val = True,
shuffle=args.shuffle,
input_size=args.input_size
)
print("load dataset done")
misc.eval_model(model, val_ds_tmp, ngpu=1, n_sample=args.n_sample)
print("======================================================")
# eval model
print("===================eval model=========================")
print(model)
val_ds = ds_fetcher(batch_size=args.batch_size,
data_root=args.data_root,
train=False,
val = True,
shuffle=args.shuffle,
input_size=args.input_size)
if args.test:
acc1, acc5 = misc.eval_model(model, val_ds, ngpu=args.ngpu, n_sample=1)
else:
acc1, acc5 = misc.eval_model(model, val_ds, ngpu=args.ngpu)
print("======================================================")
res_str = "type={}, quant_method={}, \n \
param_bits={}, bn_bits={}, fwd_bits={}, overflow_rate={},\n \
acc1={:.4f}, acc5={:.4f}".format(
args.type, args.quant_method, args.param_bits, args.bn_bits,
args.fwd_bits, args.overflow_rate, acc1, acc5)
print(res_str)
with open('acc1_acc5.txt', 'a') as f:
f.write(res_str + '\n')
# show data distribution
if args.test:
#import visdom
import numpy as np
#viz = visdom.Visdom()
import matplotlib.pyplot as plt
'''
# plot bar
for k,v in quant.extractor.items():
#print(k)
des_v= np.sort(np.abs(v).reshape(-1))
nums,times = np.unique(des_v,return_counts=True)
fig = plt.figure()
ax1 = plt.subplot(111)
width = 0.2
print(nums)
print(times)
rect = ax1.bar(left=nums,height=times,width=width,color="blue")
ax1.set_title(k)
plt.show()
'''
'''
# plot hist
for k,v in quant.extractor.items():
#print(k)
des_v= np.sort(np.abs(v).reshape(-1))
nums,times = np.unique(des_v,return_counts=True)
fig = plt.figure()
print("nums\n{}".format(nums))
print("times\n{}".format(times))
plt.hist(des_v, bins=len(nums), density=0, facecolor="blue", edgecolor="black", alpha=0.7)
plt.xlabel("nums")
# 显示纵轴标签
plt.ylabel("times")
# 显示图标题
plt.title("nums hist")
plt.show()
'''