-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathtraining.py
293 lines (232 loc) · 13.7 KB
/
training.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
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
import gc
import sys
import time
import torch
import torch.autograd
from tqdm import tqdm
from pytorch_memlab import profile
import src.utils as utils
import src.training as trainingSrc
class Trainer(trainingSrc.Trainer):
def __init__(self, params):
super().__init__()
self.fbsPruning = params.fbsPruning
def update_lr(self, params, optimiser) :
#{{{
# update learning rate
# first check if a custom scheduler is provided
if params.lrScheduler is not None:
if not hasattr(self, 'lrScheduler'):
if params.lrScheduler == 'step':
self.lrScheduler = torch.optim.lr_scheduler.StepLR(optimiser, step_size=params.stepSize, gamma=params.gamma)
else:
self.lrScheduler.step()
params.lr = self.lrScheduler.get_lr()[0]
# if not use lr schedule
elif params.lr_schedule != [] :
# get epochs to change at and lr at each of those changes
# ::2 gets every other element starting at 0
change_epochs = params.lr_schedule[::2]
new_lrs = params.lr_schedule[1::2]
epoch = params.curr_epoch
if epoch in change_epochs :
new_lr = new_lrs[change_epochs.index(epoch)]
if new_lr == -1 :
params.lr *= params.gamma
else :
params.lr = new_lr
for param_group in optimiser.param_groups :
param_group['lr'] = params.lr
return params
#}}}
def train(self, model, criterion, optimiser, inputs, targets) :
#{{{
outputs = model(inputs)
loss = criterion(outputs, targets)
if self.fbsPruning == True:
for x in model.named_buffers():
if 'g_x' in x[0]:
loss += 1e-8 * x[1]
prec1, prec5 = utils.accuracy(outputs.data, targets.data)
optimiser.zero_grad()
loss.backward()
optimiser.step()
return (loss.item(), prec1.item(), prec5.item())
#}}}
def batch_iter(self, model, criterion, optimiser, train_loader, params, losses, top1, top5):
#{{{
model.train()
for batch_idx, (inputs, targets) in tqdm(enumerate(train_loader), total=len(train_loader)-1, desc='epoch', leave=False):
# move inputs and targets to GPU
device = 'cuda:'+str(params.gpuList[0])
if params.use_cuda:
inputs, targets = inputs.cuda(device, non_blocking=True), targets.cuda(device, non_blocking=True)
# train model
loss, prec1, prec5 = self.train(model, criterion, optimiser, inputs, targets)
losses.update(loss)
top1.update(prec1)
top5.update(prec5)
if params.batchLim != -1 and batch_idx == params.batchLim:
return
#}}}
def static_finetune_l1_weights(self, params, pruner, checkpointer, train_loader, test_loader, valLoader, model, criterion, optimiser, inferer):
#{{{
print('Epoch,\tLR,\tTrain_Loss,\tTrain_Top1,\tTrain_Top5,\tTest_Loss,\tTest_Top1,\tTest_Top5,\tVal_Loss,\tVal_Top1,\tVal_Top5')
for epoch in tqdm(range(params.start_epoch, params.finetuneBudget), desc='training', leave=False) :
params.curr_epoch = epoch
state = self.update_lr(params, optimiser)
# perform pruning
if params.pruneFilters == True and epoch == params.pruneAfter:
checkpointer.save_model_only(model.state_dict(), params.printOnly, 'pre_pruning')
tqdm.write('Pruning Network')
channelsPruned, model, optimiser = pruner.prune_model(model)
totalPrunedPerc, _, _ = pruner.prune_rate(model)
tqdm.write('Pruned Percentage = {:.2f}%'.format(totalPrunedPerc))
summary = pruner.log_pruned_channels(checkpointer.root, params, totalPrunedPerc, channelsPruned)
losses = utils.AverageMeter()
top1 = utils.AverageMeter()
top5 = utils.AverageMeter()
self.batch_iter(model, criterion, optimiser, train_loader, params, losses, top1, top5)
params.train_loss = losses.avg
params.train_top1 = top1.avg
params.train_top5 = top5.avg
# get test loss
params.test_loss, params.test_top1, params.test_top5 = inferer.test_network(params, test_loader, model, criterion, optimiser, verbose=False)
params.val_loss, params.val_top1, params.val_top5 = inferer.test_network(params, valLoader, model, criterion, optimiser, verbose=False)
checkpointer.save_checkpoint(model.state_dict(), optimiser.state_dict(), params)
tqdm.write("{},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f}".format(epoch, params.lr, params.train_loss, params.train_top1, params.train_top5, params.test_loss, params.test_top1, params.test_top5, params.val_loss, params.val_top1, params.val_top5))
#}}}
def perform_pruning(self, epoch, params):
#{{{
if self.networkPruned:
return False
if epoch == 15:
return True
if epoch < 2:
self.overfitCount = 0
elif epoch == 2:
self.overfitCount = 0
self.diff = params.val_loss - params.train_loss
else:
currDiff = params.val_loss - params.train_loss
threshold = self.diff
if currDiff > threshold:
self.overfitCount += 1
else:
self.overfitCount = 0
tqdm.write("diff_thresh {:10.5f}, curr_diff {:10.5f}, oc {}".format(threshold, currDiff, self.overfitCount))
if self.overfitCount == 3:
return True
else:
return False
#}}}
def validation_finetune_l1_weights(self, params, pruner, checkpointer, train_loader, test_loader, valLoader, model, criterion, optimiser, inferer):
#{{{
print('Epoch,\tLR,\tTrain_Loss,\tTrain_Top1,\tTrain_Top5,\tTest_Loss,\tTest_Top1,\tTest_Top5,\tVal_Loss,\tVal_Top1,\tVal_Top5')
self.networkPruned = False
totalFinetuneEpochs = params.epochs
for epoch in tqdm(range(params.start_epoch, params.epochs), desc='training', leave=False) :
params.curr_epoch = epoch
if epoch == totalFinetuneEpochs:
break
# perform pruning
if self.perform_pruning(epoch, params):
tqdm.write('Pruning Network')
channelsPruned, model, optimiser = pruner.prune_model(model)
totalPrunedPerc, _, _ = pruner.prune_rate(model)
tqdm.write('Pruned Percentage = {:.2f}%'.format(totalPrunedPerc))
summary = pruner.log_pruned_channels(checkpointer.root, params, totalPrunedPerc, channelsPruned)
self.networkPruned = True
# update lr-schedule with epoch at which pruning occured before calling update_lr
lrChangeInterval = params.finetuneBudget / 3
params.lr_schedule[2] = epoch
if len(params.lr_schedule) > 4:
params.lr_schedule[4] = int(epoch + lrChangeInterval)
params.lr_schedule[6] = int(epoch + 2 * lrChangeInterval)
totalFinetuneEpochs = int(epoch + params.finetuneBudget)
tqdm.write(" ".join([str(x) for x in params.lr_schedule]))
state = self.update_lr(params, optimiser)
losses = utils.AverageMeter()
top1 = utils.AverageMeter()
top5 = utils.AverageMeter()
self.batch_iter(model, criterion, optimiser, train_loader, params, losses, top1, top5)
params.train_loss = losses.avg
params.train_top1 = top1.avg
params.train_top5 = top5.avg
# get test loss
params.test_loss, params.test_top1, params.test_top5 = inferer.test_network(params, test_loader, model, criterion, optimiser, verbose=False)
params.val_loss, params.val_top1, params.val_top5 = inferer.test_network(params, valLoader, model, criterion, optimiser, verbose=False)
checkpointer.save_checkpoint(model.state_dict(), optimiser.state_dict(), params)
tqdm.write("{},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f}".format(epoch, params.lr, params.train_loss, params.train_top1, params.train_top5, params.test_loss, params.test_top1, params.test_top5, params.val_loss, params.val_top1, params.val_top5))
#}}}
def finetune_entropy(self, params, pruner, checkpointer, train_loader, test_loader, valLoader, model, criterion, optimiser, inferer):
#{{{
print('Epoch,\tLR,\tTrain_Loss,\tTrain_Top1,\tTrain_Top5,\tTest_Loss,\tTest_Top1,\tTest_Top5,\tVal_Loss,\tVal_Top1,\tVal_Top5')
for epoch in tqdm(range(params.start_epoch, params.finetuneBudget), desc='training', leave=False) :
params.curr_epoch = epoch
state = self.update_lr(params, optimiser)
losses = utils.AverageMeter()
top1 = utils.AverageMeter()
top5 = utils.AverageMeter()
self.batch_iter(model, criterion, optimiser, train_loader, params, losses, top1, top5)
params.train_loss = losses.avg
params.train_top1 = top1.avg
params.train_top5 = top5.avg
# get test loss
params.test_loss, params.test_top1, params.test_top5 = inferer.test_network(params, test_loader, model, criterion, optimiser, verbose=False)
params.val_loss, params.val_top1, params.val_top5 = inferer.test_network(params, valLoader, model, criterion, optimiser, verbose=False)
checkpointer.save_checkpoint(model.state_dict(), optimiser.state_dict(), params)
tqdm.write("{},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f}".format(epoch, params.lr, params.train_loss, params.train_top1, params.train_top5, params.test_loss, params.test_top1, params.test_top5, params.val_loss, params.val_top1, params.val_top5))
#}}}
def single_forward_backward(self, params, model, criterion, optimiser, train_loader):
#{{{
for batch_idx, (inputs, targets) in enumerate(train_loader):
if params.use_cuda :
inputs, targets = inputs.cuda(), targets.cuda()
loss, prec1, prec5 = self.train(model, criterion, optimiser, inputs, targets)
return
#}}}
#{{{
# def finetune_newtork(self, params, pruner, checkpointer, train_loader, test_loader, valLoader, model, criterion, optimiser, inferer):
# print('Epoch,\tLR,\tTrain_Loss,\tTrain_Top1,\tTrain_Top5,\tTest_Loss,\tTest_Top1,\tTest_Top5,\tVal_Loss,\tVal_Top1,\tVal_Top5')
#
# for epoch in tqdm(range(params.start_epoch, params.epochs), desc='training', leave=False) :
# params.curr_epoch = epoch
# state = self.update_lr(params, optimiser)
#
# losses = utils.AverageMeter()
# top1 = utils.AverageMeter()
# top5 = utils.AverageMeter()
# self.batch_iter(model, criterion, optimiser, train_loader, params, losses, top1, top5)
# params.train_loss = losses.avg
# params.train_top1 = top1.avg
# params.train_top5 = top5.avg
#
# # perform pruning
# # if (params.pruneWeights == True or params.pruneFilters == True) and ((epoch+1) % params.pruneAfter == 0):
# if (params.pruneWeights == True or params.pruneFilters == True) and (epoch = params.finetuneBudget):
# tqdm.write('Pruning Network')
# model = pruner.prune_model(model)
# params.pruningPerc += params.prunePercIncrement
# totalPrunedPerc = pruner.prune_rate(model, True)
# tqdm.write('Pruned Percentage = {}'.format(totalPrunedPerc))
# # checkpointer.log_prune_rate(params, totalPrunedPerc)
# pruner.log_prune_rate(checkpointer.root, params, totalPrunedPerc)
# params.prunePercPerLayer = []
# if params.fbsPruning == True and ((epoch + 1) % params.pruneAfter == 0):
# params.unprunedRatio -= (params.prunePercIncrement / 100.0)
# # return if pruning become lower than lower bound
# if params.unprunedRatio <= params.unprunedLB:
# return
#
# tqdm.write('Pruning Network with FBS')
# model = pruner.prune_model(model)
# tqdm.write('Pruned Percentage = {}'.format(1.0 - params.unprunedRatio))
#
# # get test loss
# params.test_loss, params.test_top1, params.test_top5 = inferer.test_network(params, test_loader, model, criterion, optimiser)
# params.val_loss, params.val_top1, params.val_top5 = inferer.test_network(params, valLoader, model, criterion, optimiser)
# checkpointer.save_checkpoint(model.state_dict(), optimiser.state_dict(), params)
#
# tqdm.write("{},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f},\t{:10.5f}".format(epoch, params.lr, params.train_loss, params.train_top1, params.train_top5, params.test_loss, params.test_top1, params.test_top5, params.val_loss, params.val_top1, params.val_top5))
#}}}