print "If you have used Dynamic Model, make sure you pass correct parameters"
raise SystemExit
#fit the model
lrmodel.fit(train_x,train_y,batch_size=batchsize,epochs=epochs,verbose=1)
#make prediction
pred=lrmodel.predict(test_x, batch_size=32)
pred = [ii.argmax()for ii in pred]
test_y = [ii.argmax()for ii in test_y]
results.append(accuracy_score(pred,test_y))
print accuracy_score(pred,test_y)
jj=str(set(list(test_y)))
print "Unique in test_y",jj
print "Results: " + str( np.array(results).mean() )
else:
train_x=tr_X
train_y=np.array(tr_y)
print "Evaluation mode"
lrmodel=miz.prepare_model()
train_y = to_categorical(train_y,num_classes=len(labels))
#fit the model
lrmodel.fit(train_x,train_y,batch_size=64,epochs=50,verbose=1)
truth,pred=test(lrmodel,txt_eva_path,new_p,model)
acc=aud_utils.calculate_accuracy(truth,pred)
print "Accuracy %.2f prcnt"%acc
lrmodel.fit(train_x,train_y,batch_size=batchsize,epochs=epochs,verbose=1)
#make prediction
pred=lrmodel.predict(test_x, batch_size=32)
pred = [ii.argmax()for ii in pred]
test_y = [ii.argmax()for ii in test_y]
results.append(accuracy_score(pred,test_y))
print accuracy_score(pred,test_y)
jj=str(set(list(test_y)))
print "Unique in test_y",jj
print "Results: " + str( np.array(results).mean() )
else:
train_x=np.array(tr_X)
train_y=np.array(tr_y)
print "Evaluation mode"
lrmodel=miz.prepare_model()
train_y = to_categorical(train_y,num_classes=len(labels))
#fit the model
lrmodel.fit(train_x,train_y,batch_size=batchsize,epochs=epochs,verbose=1)
truth,pred=test(lrmodel,txt_eva_path)
from sklearn.metrics import accuracy_score
acc1=accuracy_score(truth, pred)
print acc1
acc=aud_utils.calculate_accuracy(truth.sort(),pred.sort())
print "Accuracy %.2f prcnt"%acc
