# contextLSTMCNN


###########expADE.py##########
Usage: expADE.py [options]

Options:
  -h, --help            show this help message and exit
  --txtFile=TXTFILE     txt input file
  --pklFile=PKLFILE     pickled input file
  --train=TRAIN         train model
  --test                test model
  --trainTestSplit      split file into training (80%) and testing parts(20%)
  --saveTrainTest=SAVETRAINTEST
                        save train test split
  --saveModel=SAVEMODEL
                        save trained model
  --epochs=EPOCHS       number of epochs for training, default 50
  --testIDs=TESTIDS     load test IDs
  --trainIDs=TRAINIDS   load train IDs
  --windowSize=WINDOWSSIZE
                        window size
  --loadModel=LOADMODEL
                        load trianed model
  --nFoldsSplit=NFOLDSSPLIT
                        prefix of n folds split
  --sentenceLevelEvaluation
                        sentence level evaluation
  --experiment          compare with other models
  --sentAlpha=SENTALPHA
                        sent alpha
  --blstmBlstmcnn       train blstmBlstmcnn
  --loadW2V=LOADW2V     load trianed w2v
  --wordAlpha=WORDALPHA
                        word alpha

############expIEMOCAP.py############
Usage: expIEMOCAP.py [options]

Options:
  -h, --help            show this help message and exit
  --transcriptions=TRANSCRIPTIONDIR
                        transcription directory
  --evaluation=EVALUATIONDIR
                        evaluation directory
  --train=TRAIN         train model
  --saveData            train model
  --test                test model
  --trainTestSplit      split file into training (80%) and testing parts(20%)
  --saveTrainTest=SAVETRAINTEST
                        save train test split
  --nFoldsSplit=NFOLDSSPLIT
                        prefix of n folds split
  --saveModel=SAVEMODEL
                        save trained model
  --epochs=EPOCHS       number of epochs for training, default 50
  --testIDs=TESTIDS     load test IDs
  --trainIDs=TRAINIDS   load train IDs
  --windowSize=WINDOWSSIZE
                        window size
  --loadModel=LOADMODEL
                        load trianed model
  --sentAlpha=SENTALPHA
                        sent alpha
  --wordAlpha=WORDALPHA
                        word alpha
  --experiment          compare with other models
  --blstmBlstmcnn       train blstmBlstmcnn
  --loadW2V=LOADW2V     load trianed w2v

Cite:
Xingyi Song, Johann Petrak, Angus Roberts. A Deep Neural Network Sentence Level
Classification Method with Context Information. In: EMNLP, 2018