def train(**kwargs):
conf.parse(kwargs)
# train_set = DataSet(cfg, train=True, test=False)
train_set = ImageFolder(conf.TRAIN_DATA_ROOT, transform)
train_loader = DataLoader(train_set, conf.BATCH_SIZE,
shuffle=True,
num_workers=conf.NUM_WORKERS)
model = Network()
if conf.LOAD_MODEL_PATH:
print(conf.LOAD_MODEL_PATH)
model.load_state_dict(torch.load(conf.CHECKPOINTS_ROOT + conf.LOAD_MODEL_PATH))
device = torch.device('cuda:0' if conf.USE_GPU else 'cpu')
criterion = nn.CrossEntropyLoss().to(device)
lr = conf.LEARNING_RATE
optim = torch.optim.Adam(params=model.parameters(),
lr=lr,
weight_decay=conf.WEIGHT_DECAY)
model.to(device)
for epoch in range(conf.MAX_EPOCH):
model.train()
running_loss = 0
for step, (inputs, targets) in tqdm(enumerate(train_loader)):
inputs, targets = inputs.to(device), targets.to(device)
optim.zero_grad()
outs = model(inputs)
loss = criterion(outs, targets)
loss.backward()
optim.step()
running_loss += loss.item()
if step % conf.PRINT_FREQ == conf.PRINT_FREQ - 1:
running_loss = running_loss / conf.PRINT_FREQ
print('[%d, %5d] loss: %.3f' % (epoch + 1, step + 1, running_loss))
# vis.plot('loss', running_loss)
running_loss = 0
torch.save(model.state_dict(), conf.CHECKPOINTS_ROOT + time.strftime('%Y-%m-%d-%H-%M-%S.pth'))
for param_group in optim.param_groups:
lr *= conf.LEARNING_RATE_DECAY
param_group['lr'] = lr
# -*- coding: utf-8 -*-
import os
from config import Config
from model.data import Data
from model.network import Network
if __name__ == '__main__':
dir_path = os.path.join(os.path.dirname(os.path.realpath(__file__)),
Config.CURRENT_MODEL_BASE_PATH)
data = Data()
data.load_data_from_file(os.path.join(dir_path,
'data.nosync/all_data.npy'))
model = Network(data, os.path.join(dir_path, 'log.nosync/network/run1'))
model.train()
class LogisticRegression(Classifier):
"""
A digit-7 recognizer based on logistic regression algorithm
Parameters
----------
train : list
valid : list
test : list
learningRate : float
epochs : positive int
Attributes
----------
trainingSet : list
validationSet : list
testSet : list
weight : list
learningRate : float
epochs : positive int
"""
def __init__(self, train, valid, test, learningRate=0.1, epochs=50):
self.learningRate = learningRate
self.epochs = epochs
self.trainingSet = train
self.validationSet = valid
self.testSet = test
self.network = Network(learningRate)
def train(self, verbose=True):
"""Train the Logistic Regression.
Parameters
----------
verbose : boolean
Print logging messages with validation accuracy if verbose is True.
"""
# Here you have to implement training method "epochs" times
# Please using LogisticLayer class
self.epochs = 1
for i in range(0,self.epochs):
print i
self.network.train(self.trainingSet.input, self.trainingSet.label)
pass
def classify(self, testInstance):
"""Classify a single instance.
Parameters
----------
testInstance : list of floats
Returns
-------
bool :
True if the testInstance is recognized as a 7, False otherwise.
"""
# Here you have to implement classification method given an
# instance
res = self.network.classify(testInstance)
if res > 0.5:
return 1
else:
return 0
def evaluate(self, test=None):
"""Evaluate a whole dataset.
Parameters
----------
test : the dataset to be classified
if no test data, the test set associated to the classifier will be used
Returns
-------
List:
List of classified decisions for the dataset's entries.
"""
if test is None:
test = self.testSet.input
# Once you can classify an instance, just use map for all of the test
# set.
return list(map(self.classify, test))
