def train(self):
outputPrefix=self.readField(self.config,self.name,"output_directory")
outputDir=os.path.join(outputPrefix,self.name)
if not os.path.exists(outputDir):
os.mkdir(outputDir)
imageinput = self.readField(self.isae.ae[1].config, self.isae.ae[1].name, "train_data")
textinput = self.readField(self.tsae.ae[1].config, self.tsae.ae[1].name, "train_data")
if self.readField(self.config, self.name,"extract_reps")=="True":
imageoutput=self.readField(self.isae.ae[-1].config, self.isae.ae[-1].name, "train_reps")
textoutput=self.readField(self.tsae.ae[-1].config, self.tsae.ae[-1].name, "train_reps")
else:
imageoutput=None
textoutput=None
maxEpoch = int(self.readField(self.config, self.name, "max_epoch"))
trainSize=int(self.readField(self.config, self.name, "train_size"))
numBatch = int(trainSize / self.batchsize)
normalizeImg=self.str2bool(self.readField(self.config, self.name, "normalize"))
imgTrainDH=DataHandler(imageinput, imageoutput, self.isae.ae[1].vDim, self.isae.ae[-1].hDim, self.batchsize, numBatch,normalizeImg)
txtTrainDH=DataHandler(textinput, textoutput, self.tsae.ae[1].vDim, self.tsae.ae[-1].hDim, self.batchsize, numBatch)
showFreq = int(self.readField(self.config, self.name, "show_freq"))
if showFreq > 0:
visDir = os.path.join(outputDir, "vis")
if not os.path.exists(visDir):
os.makedirs(visDir)
evalFreq = int(self.readField(self.config, self.name, "eval_freq"))
if evalFreq!=0:
qsize=int(self.readField(self.config, self.name, "query_size"))
labelPath=self.readField(self.config,self.name,"label")
label=np.load(labelPath)
queryPath=self.readField(self.config, self.name, "query")
validation=evaluate.Evaluator(queryPath,label,os.path.join(outputDir,'perf'), self.name, query_size=qsize,verbose=self.verbose)
validateImagepath = self.readField(self.isae.ae[1].config, self.isae.ae[1].name, "validation_data")
validateTextpath = self.readField(self.tsae.ae[1].config, self.tsae.ae[1].name, "validation_data")
validateImgData = gp.garray(np.load(validateImagepath))
if normalizeImg:
validateImgData=imgTrainDH.doNormalization(validateImgData)
validateTxtData = gp.garray(np.load(validateTextpath))
else:
print "Warning: no evluation setting!"
nCommon, nMetric, title=self.getDisplayFields()
if self.verbose:
print title
for epoch in range(maxEpoch):
perf=np.zeros( nMetric)
epoch1, imgcost, txtcost, diffcost=self.checkPath(epoch)
imgTrainDH.reset()
txtTrainDH.reset()
for i in range(numBatch):
img = imgTrainDH.getOneBatch()
txt = txtTrainDH.getOneBatch()
curr= self.trainOneBatch(img, txt, epoch1, imgcost, txtcost, diffcost)
perf=self.aggregatePerf(perf, curr)
if evalFreq!=0 and (1+epoch) % evalFreq == 0:
imgcode,txtcode=self.getReps(validateImgData, validateTxtData)
validation.evalCrossModal(imgcode,txtcode,epoch,'V')
if showFreq != 0 and (1+epoch) % showFreq == 0:
imgcode,txtcode=self.getReps(validateImgData, validateTxtData)
np.save(os.path.join(visDir,'%simg' % str((epoch+1)/showFreq)),imgcode)
np.save(os.path.join(visDir,'%stxt' % str((epoch+1)/showFreq)),txtcode)
if self.verbose:
self.printEpochInfo(epoch, perf, nCommon)
if self.readField(self.config, self.name, "checkpoint")=="True":
self.doCheckpoint(outputDir)
if self.readField(self.config, self.name,"extract_reps")=="True":
if evalFreq!=0:
self.extractValidationReps(validateImgData, validateTxtData, "validation_data","validation_reps")
self.extractTrainReps(imgTrainDH, txtTrainDH, numBatch)
self.saveConfig(outputDir)
def train(self):
outputPrefix = self.readField(self.config, self.name,
"output_directory")
outputDir = os.path.join(outputPrefix, self.name)
if not os.path.exists(outputDir):
os.makedirs(outputDir)
showFreq = int(self.readField(self.config, self.name, "show_freq"))
if showFreq > 0:
visDir = os.path.join(outputDir, 'vis')
if not os.path.exists(visDir):
os.mkdir(visDir)
#do normalization for images if they are not normalized before
normalize = self.str2bool(
self.readField(self.config, self.name, "normalize"))
trainDataSize = int(
self.readField(self.config, self.name, "train_size"))
numBatch = trainDataSize / self.batchsize
trainDataPath = self.readField(self.config, self.name, "train_data")
if self.readField(self.config, self.name, "extract_reps") == "True":
trainRepsPath = self.readField(self.config, self.name,
"train_reps")
else:
trainRepsPath = None
trainDataLoader = DataHandler(trainDataPath, trainRepsPath, self.vDim,
self.hDim, self.batchsize, numBatch,
normalize)
evalFreq = int(self.readField(self.config, self.name, 'eval_freq'))
if evalFreq != 0:
qsize = int(self.readField(self.config, self.name, "query_size"))
evalPath = self.readField(self.config, self.name,
"validation_data")
labelPath = self.readField(self.config, self.name, "label")
queryPath = self.readField(self.config, self.name, "query")
label = np.load(labelPath)
eval = Evaluator(queryPath,
label,
os.path.join(outputDir, 'perf'),
self.name,
query_size=qsize,
verbose=self.verbose)
validation_data = gp.garray(np.load(evalPath))
if normalize:
validation_data = trainDataLoader.doNormalization(
validation_data)
maxEpoch = int(self.readField(self.config, self.name, "max_epoch"))
nCommon, nMetric, title = self.getDisplayFields()
if self.verbose:
print title
for epoch in range(maxEpoch):
perf = np.zeros(nMetric)
trainDataLoader.reset()
for i in range(numBatch):
batch = trainDataLoader.getOneBatch()
curr = self.trainOneBatch(batch, epoch, computeStat=True)
perf = self.aggregatePerf(perf, curr)
if showFreq != 0 and (1 + epoch) % showFreq == 0:
validation_code = self.getReps(validation_data)
np.save(os.path.join(visDir, '%dvis' % (1 + epoch)),
validation_code)
if evalFreq != 0 and (1 + epoch) % evalFreq == 0:
validation_code = self.getReps(validation_data)
eval.evalSingleModal(validation_code, epoch, self.name + 'V')
validation_code = None
if self.verbose:
self.printEpochInfo(epoch, perf, nCommon)
if self.readField(self.config, self.name, "checkpoint") == "True":
self.doCheckpoint(outputDir)
if self.readField(self.config, self.name, "extract_reps") == "True":
if evalFreq != 0:
validation_reps_path = self.readField(self.config, self.name,
"validation_reps")
self.extractValidationReps(validation_data,
validation_reps_path)
self.extractTrainReps(trainDataLoader, numBatch)
self.saveConfig(outputDir)
def train(self):
outputPrefix = self.readField(self.config, self.name,
"output_directory")
outputDir = os.path.join(outputPrefix, self.name)
if not os.path.exists(outputDir):
os.mkdir(outputDir)
imageinput = self.readField(self.isae.ae[1].config,
self.isae.ae[1].name, "train_data")
textinput = self.readField(self.tsae.ae[1].config,
self.tsae.ae[1].name, "train_data")
if self.readField(self.config, self.name, "extract_reps") == "True":
imageoutput = self.readField(self.isae.ae[-1].config,
self.isae.ae[-1].name, "train_reps")
textoutput = self.readField(self.tsae.ae[-1].config,
self.tsae.ae[-1].name, "train_reps")
else:
imageoutput = None
textoutput = None
maxEpoch = int(self.readField(self.config, self.name, "max_epoch"))
trainSize = int(self.readField(self.config, self.name, "train_size"))
numBatch = int(trainSize / self.batchsize)
normalizeImg = self.str2bool(
self.readField(self.config, self.name, "normalize"))
imgTrainDH = DataHandler(imageinput, imageoutput, self.isae.ae[1].vDim,
self.isae.ae[-1].hDim, self.batchsize,
numBatch, normalizeImg)
txtTrainDH = DataHandler(textinput, textoutput, self.tsae.ae[1].vDim,
self.tsae.ae[-1].hDim, self.batchsize,
numBatch)
showFreq = int(self.readField(self.config, self.name, "show_freq"))
if showFreq > 0:
visDir = os.path.join(outputDir, "vis")
if not os.path.exists(visDir):
os.makedirs(visDir)
evalFreq = int(self.readField(self.config, self.name, "eval_freq"))
if evalFreq != 0:
qsize = int(self.readField(self.config, self.name, "query_size"))
labelPath = self.readField(self.config, self.name, "label")
label = np.load(labelPath)
queryPath = self.readField(self.config, self.name, "query")
validation = evaluate.Evaluator(queryPath,
label,
os.path.join(outputDir, 'perf'),
self.name,
query_size=qsize,
verbose=self.verbose)
validateImagepath = self.readField(self.isae.ae[1].config,
self.isae.ae[1].name,
"validation_data")
validateTextpath = self.readField(self.tsae.ae[1].config,
self.tsae.ae[1].name,
"validation_data")
validateImgData = gp.garray(np.load(validateImagepath))
if normalizeImg:
validateImgData = imgTrainDH.doNormalization(validateImgData)
validateTxtData = gp.garray(np.load(validateTextpath))
else:
print "Warning: no evluation setting!"
nCommon, nMetric, title = self.getDisplayFields()
if self.verbose:
print title
for epoch in range(maxEpoch):
perf = np.zeros(nMetric)
epoch1, imgcost, txtcost, diffcost = self.checkPath(epoch)
imgTrainDH.reset()
txtTrainDH.reset()
for i in range(numBatch):
img = imgTrainDH.getOneBatch()
txt = txtTrainDH.getOneBatch()
curr = self.trainOneBatch(img, txt, epoch1, imgcost, txtcost,
diffcost)
perf = self.aggregatePerf(perf, curr)
if evalFreq != 0 and (1 + epoch) % evalFreq == 0:
imgcode, txtcode = self.getReps(validateImgData,
validateTxtData)
validation.evalCrossModal(imgcode, txtcode, epoch, 'V')
if showFreq != 0 and (1 + epoch) % showFreq == 0:
imgcode, txtcode = self.getReps(validateImgData,
validateTxtData)
np.save(
os.path.join(visDir, '%simg' % str(
(epoch + 1) / showFreq)), imgcode)
np.save(
os.path.join(visDir, '%stxt' % str(
(epoch + 1) / showFreq)), txtcode)
if self.verbose:
self.printEpochInfo(epoch, perf, nCommon)
if self.readField(self.config, self.name, "checkpoint") == "True":
self.doCheckpoint(outputDir)
if self.readField(self.config, self.name, "extract_reps") == "True":
if evalFreq != 0:
self.extractValidationReps(validateImgData, validateTxtData,
"validation_data",
"validation_reps")
self.extractTrainReps(imgTrainDH, txtTrainDH, numBatch)
self.saveConfig(outputDir)
def train(self):
outputPrefix=self.readField(self.config,self.name,"output_directory")
outputDir=os.path.join(outputPrefix,self.name)
if not os.path.exists(outputDir):
os.makedirs(outputDir)
showFreq = int(self.readField(self.config, self.name, "show_freq"))
if showFreq > 0:
visDir = os.path.join(outputDir,'vis')
if not os.path.exists(visDir):
os.mkdir(visDir)
#do normalization for images if they are not normalized before
normalize=self.str2bool(self.readField(self.config, self.name, "normalize"))
trainDataSize=int(self.readField(self.config, self.name, "train_size"))
numBatch = trainDataSize / self.batchsize
trainDataPath = self.readField(self.config, self.name, "train_data")
if self.readField(self.config,self.name,"extract_reps")=="True":
trainRepsPath=self.readField(self.config, self.name, "train_reps")
else:
trainRepsPath=None
trainDataLoader=DataHandler(trainDataPath, trainRepsPath, self.vDim, self.hDim, self.batchsize,numBatch, normalize)
evalFreq=int(self.readField(self.config,self.name,'eval_freq'))
if evalFreq!=0:
qsize=int(self.readField(self.config, self.name, "query_size"))
evalPath=self.readField(self.config,self.name,"validation_data")
labelPath=self.readField(self.config,self.name,"label")
queryPath=self.readField(self.config, self.name, "query")
label=np.load(labelPath)
eval=Evaluator(queryPath,label ,os.path.join(outputDir,'perf'), self.name, query_size=qsize,verbose=self.verbose)
validation_data=gp.garray(np.load(evalPath))
if normalize:
validation_data=trainDataLoader.doNormalization(validation_data)
maxEpoch = int(self.readField(self.config, self.name, "max_epoch"))
nCommon, nMetric, title=self.getDisplayFields()
if self.verbose:
print title
for epoch in range(maxEpoch):
perf=np.zeros( nMetric)
trainDataLoader.reset()
for i in range(numBatch):
batch = trainDataLoader.getOneBatch()
curr = self.trainOneBatch(batch, epoch, computeStat=True)
perf=self.aggregatePerf(perf, curr)
if showFreq != 0 and (1+epoch) % showFreq == 0:
validation_code=self.getReps(validation_data)
np.save(os.path.join(visDir, '%dvis' % (1+epoch)), validation_code)
if evalFreq !=0 and (1+epoch) % evalFreq ==0:
validation_code=self.getReps(validation_data)
eval.evalSingleModal(validation_code,epoch,self.name+'V')
validation_code=None
if self.verbose:
self.printEpochInfo(epoch,perf,nCommon)
if self.readField(self.config,self.name,"checkpoint")=="True":
self.doCheckpoint(outputDir)
if self.readField(self.config,self.name,"extract_reps")=="True":
if evalFreq!=0:
validation_reps_path=self.readField(self.config, self.name, "validation_reps")
self.extractValidationReps(validation_data, validation_reps_path)
self.extractTrainReps(trainDataLoader, numBatch)
self.saveConfig(outputDir)