bitprune/logger.py at main · DensoITLab/bitprune · GitHub

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
from easydict import EasyDict as edict
import numpy as np
import pickle
import os
import torch

class AverageMeter(object):
    """Computes and stores the average and current value"""
    def __init__(self):
        self.initialized = False
        self.val = None
        self.avg = None
        self.sum = None
        self.count = None

    def initialize(self, val, weight):
        self.val = val
        self.avg = val
        self.sum = val * weight
        self.count = weight
        self.initialized = True

    def update(self, val, weight=1):
        if not self.initialized:
            self.initialize(val, weight)
        else:
            self.add(val, weight)

    def add(self, val, weight):
        self.val = val
        self.sum += val * weight
        self.count += weight
        self.avg = self.sum / self.count

    def value(self):
        return self.val

    def average(self):
        return self.avg

class STATS():
    def __init__(self, config):
        self.cfg = config
        log = edict()
        log.train =  edict()
        log.val =  edict()
        log.test =  edict()


        def init_loss(e_obj):
            e_obj.ce_loss         = np.zeros(config.optim.epochs)
            e_obj.acc             = np.zeros(config.optim.epochs)
            e_obj.acc5            = np.zeros(config.optim.epochs)
            e_obj.mac             = np.zeros(config.optim.epochs)
            e_obj.sac             = np.zeros(config.optim.epochs)
            e_obj.mac2            = np.zeros(config.optim.epochs)
            e_obj.sac2            = np.zeros(config.optim.epochs)
            e_obj.op_loss         = np.zeros(config.optim.epochs)
            e_obj.avg_wgt         = np.zeros(config.optim.epochs)
            e_obj.avg_act         = np.zeros(config.optim.epochs)
            e_obj.avg_bit         = np.zeros(config.optim.epochs)
            e_obj.lr              = np.zeros(config.optim.epochs)
        init_loss(log.train)
        init_loss(log.val)
        init_loss(log.test)

        log.best = edict()
        log.best.acc = 0.0
        log.best.ce_loss = 0.0
        log.best.itr = 0
        log.best.epoch = -1

        self.log = log
        self.epoch = 0

    def init_meter(self, phase):
        self.phase = phase
        if phase=='train':
            print("<<<<<<<<< start epoch{:3d} {} {}".format(self.epoch,  self.cfg.dataset.name, self.cfg.cfg_name + ' ' +self.cfg.file_name))

        self.ce_loss_meter      = AverageMeter()
        self.acc_meter          = AverageMeter()
        self.acc5_meter         = AverageMeter()
        self.mac_meter          = AverageMeter()
        self.sac_meter          = AverageMeter()
        self.mac2_meter         = AverageMeter()
        self.sac2_meter         = AverageMeter()
        self.op_loss_meter      = AverageMeter()
        self.avg_wgt_meter      = AverageMeter()
        self.avg_act_meter      = AverageMeter()
        self.avg_bit_meter      = AverageMeter()
        self.ce_loss_meter      = AverageMeter()
        self.lr_meter           = AverageMeter()

    def update_meter(self, ce_loss, outputs, labels, lr, layer_stats):
        acc1, acc5 = self.__accuracy(outputs, labels, topk=(1, 5))

        self.ce_loss_meter.update(ce_loss.item())
        self.acc_meter.update(acc1.item())
        self.acc5_meter.update(acc5.item())
        self.op_loss_meter.update(layer_stats.op_loss.item())
        self.mac_meter.update(layer_stats.mac)
        self.sac_meter.update(layer_stats.sac)
        self.mac2_meter.update(layer_stats.mac2)
        self.sac2_meter.update(layer_stats.sac2)
        self.avg_wgt_meter.update(layer_stats.avg_wgt)
        self.avg_act_meter.update(layer_stats.avg_act)
        self.avg_bit_meter.update(layer_stats.avg_bit)
        self.lr_meter.update(lr)


    def save(self, logname='log2'):
        self.log[self.phase].ce_loss[self.epoch]    = self.ce_loss_meter.average()
        self.log[self.phase].acc[self.epoch]        = self.acc_meter.average()
        self.log[self.phase].acc5[self.epoch]       = self.acc5_meter.average()
        self.log[self.phase].mac[self.epoch]        = self.mac_meter.average()
        self.log[self.phase].sac[self.epoch]        = self.sac_meter.average()
        self.log[self.phase].mac2[self.epoch]       = self.mac2_meter.average()
        self.log[self.phase].sac2[self.epoch]       = self.sac2_meter.average()
        self.log[self.phase].op_loss[self.epoch]    = self.op_loss_meter.average()
        self.log[self.phase].avg_wgt[self.epoch]    = self.avg_wgt_meter.average()
        self.log[self.phase].avg_act[self.epoch]    = self.avg_act_meter.average()
        self.log[self.phase].avg_bit[self.epoch]    = self.avg_bit_meter.average()
        self.log[self.phase].lr[self.epoch]         = self.lr_meter.average()


        if self.phase=='val' and self.log[self.phase].acc[self.epoch]>self.log.best.acc:
            self.log.best.acc = self.log[self.phase].acc[self.epoch]
            self.log.best.epoch = self.epoch
            print(" Best updated!  epoch{:3d}, acc: {:0.3f}, ".format(self.log.best.epoch, self.log.best.acc))

        with open(os.path.join(self.cfg.stats_dir, logname), 'wb') as handle:
            pickle.dump(self.log, handle, protocol=pickle.HIGHEST_PROTOCOL)

        # print("finished epoch{:3d} {}>>>>>>>>>>".format(self.epoch, self.phase,))

    def disp(self):
        print_scale_op_loss    = 1e-3
        print_scale_mac        = 1e9

        print(" {}, lr:{:0.5f} epoch{:3d}, acc: {:0.3f}, acc5: {:0.3f}, ce: {:2.3f}, spr: {:2.3f},  mac: {:2.3f}, {:2.3f}, sac: {:2.3f},{:2.3f}  avg_bit: {:0.4f}, avg_wgt: {:0.4f}, avg_act: {:0.4f}".format(
            self.phase,
            self.lr_meter.average(),
            self.epoch,
            self.acc_meter.average(),
            self.acc5_meter.average(),
            self.ce_loss_meter.average(),
            self.op_loss_meter.average()/print_scale_op_loss,
            self.mac_meter.average()/print_scale_mac,
            self.mac2_meter.average()/print_scale_mac,
            self.sac_meter.average()/print_scale_mac,
            self.sac2_meter.average()/print_scale_mac,
            self.avg_bit_meter.average(),
            self.avg_wgt_meter.average(),
            self.avg_act_meter.average())
        )


    @staticmethod
    def __accuracy(output, target, topk=(1,)):
        """Computes the accuracy over the k top predictions for the specified values of k."""
        with torch.no_grad():
            maxk = max(topk)
            batch_size = target.size(0)

            _, pred = output.topk(maxk, 1, True, True)
            pred = pred.t()
            correct = pred.eq(target.view(1, -1).expand_as(pred))

            res = []
            for k in topk:
                correct_k = correct[:k].reshape(-1).float().sum(0, keepdim=True)
                res.append(correct_k.mul_(1.0 / batch_size))
            return res