if opt.Cross_Validation == 'k_fold':
signals, stages = transformer.batch_generator(signals,
stages,
opt.batchsize,
shuffle=True)
train_sequences, test_sequences = transformer.k_fold_generator(
len(stages), opt.fold_num)
batch_length = len(stages)
print('length of batch:', batch_length)
show_freq = int(len(train_sequences[0]) / 5)
util.show_menory()
t2 = time.time()
print('load data cost time: %.2f' % (t2 - t1), 's')
net = CreatNet(opt.model_name)
torch.save(net.cpu().state_dict(), './checkpoints/' + opt.model_name + '.pth')
weight = np.array([1, 1, 1, 1, 1])
if opt.weight_mod == 'avg_best':
weight = np.log(1 / stage_cnt_per)
weight[2] = weight[2] + 1
weight = np.clip(weight, 1, 5)
print('Loss_weight:', weight)
weight = torch.from_numpy(weight).float()
# print(net)
if not opt.no_cuda:
net.cuda()
weight = weight.cuda()
if not opt.no_cudnn:
cudnn.benchmark = True
import numpy as np
import torch
import matplotlib.pyplot as plt
import util
import transformer
import dataloader
from options import Options
from creatnet import CreatNet
'''
@hypox64
19/04/03
'''
opt = Options().getparse()
net=CreatNet(opt.model_name)
if not opt.no_cuda:
net.cuda()
if not opt.no_cudnn:
import torch.backends.cudnn as cudnn
cudnn.benchmark = True
if opt.pretrained:
net.load_state_dict(torch.load('./checkpoints/pretrained/'+opt.model_name+'.pth'))
# N3(S4+S3)->0 N2->1 N1->2 REM->3 W->4
stage_map={0:'stage3',1:'stage2',2:'stage3',3:'REM',4:'Wake'}
def runmodel(eegdata):
eegdata = eegdata.reshape(1,-1)
eegdata = transformer.ToInputShape(eegdata,opt.model_name,test_flag =True)
# 2.shape signals:[num,ch,length] labels:[num]
# num:samples_num, ch :channel_num, num:length of each sample
# for example:
signals = np.zeros((10,1,10),dtype='np.float64')
labels = np.array([0,0,0,0,0,1,1,1,1,1]) #0->class0 1->class1
* step2: input ```--dataset_dir your_dataset_dir``` when running code.
'''
signals, labels = dataloader.loaddataset(opt)
label_cnt, label_cnt_per, _ = statistics.label_statistics(labels)
train_sequences, test_sequences = transformer.k_fold_generator(
len(labels), opt.k_fold)
t2 = time.time()
print('load data cost time: %.2f' % (t2 - t1), 's')
net = CreatNet(opt)
util.writelog('network:\n' + str(net), opt, True)
util.show_paramsnumber(net, opt)
weight = np.ones(opt.label)
if opt.weight_mod == 'auto':
weight = 1 / label_cnt_per
weight = weight / np.min(weight)
util.writelog('label statistics: ' + str(label_cnt), opt, True)
util.writelog('Loss_weight:' + str(weight), opt, True)
weight = torch.from_numpy(weight).float()
# print(net)
if not opt.no_cuda:
net.cuda()
weight = weight.cuda()
import torch
import matplotlib.pyplot as plt
import util
import transformer
import dataloader
from options import Options
from creatnet import CreatNet
'''
--------------------------------preload data--------------------------------
@hypox64
2020/04/03
'''
opt = Options().getparse()
net = CreatNet(opt)
#load data
signals = np.load('./datasets/simple_test/signals.npy')
labels = np.load('./datasets/simple_test/labels.npy')
#load prtrained_model
net.load_state_dict(torch.load('./checkpoints/pretrained/micro_multi_scale_resnet_1d_50class.pth'))
net.eval()
if not opt.no_cuda:
net.cuda()
for signal,true_label in zip(signals, labels):
signal = signal.reshape(1,1,-1) #batchsize,ch,length
true_label = true_label.reshape(1) #batchsize
signal,true_label = transformer.ToTensor(signal,true_label,no_cuda =opt.no_cuda)