[KT클라우드]AI Studio 소스코드 가이드 - 샘플코드

from __future__ import print_function

import numpy as np

import pandas as pd

from tensorflow.python import keras

from tensorflow.python.keras import backend as K

from tensorflow.python.keras.models import Sequential

from tensorflow.python.keras.layers import Dense, Dropout, Flatten

from tensorflow.python.keras.layers import Conv2D, MaxPooling2D

from sklearn.datasets import load_iris

from sklearn.preprocessing import LabelEncoder

from sklearn.model_selection import train_test_split

"""

AI Studio와 연계를 위한 기본 객체 생성

"""

from aicentro.session import Session

aistudio_session = Session(verify=False)

"""

학습 모델 개발 시 프레임워크별 객체 사용

"""

from aicentro.framework.keras import Keras as AistudioFrm

aistudio_framework = AistudioFrm(session=aistudio_session)

"""

모델 코드 작성

"""

batch_size = 128

num_classes = 3

epochs = 10

data = load_iris()

X = data.data

y = data.target

encoder = LabelEncoder()

y1 = encoder.fit_transform(y)

Y = pd.get_dummies(y1).values

X_train, X_test, y_train, y_test = train_test_split(X, Y,test_size=0.2, random_state=1)

model = Sequential()

model.add(Dense(64,input_shape=(4,),activation='relu'))

model.add(Dense(64,activation='relu'))

model.add(Dense(3,activation='softmax'))

model.compile(loss='categorical_crossentropy', optimizer='Adam',metrics=['accuracy'])

model.summary()

"""

모델 학습 시 Accuracy 와 Loss 값을 AI Studio로 전송하여 UI 상에 노출

제공된 Metric Callback 사용

"""

history = aistudio_framework.get_metric_callback()

model.fit(X_train, y_train,

       batch_size=batch_size,

       epochs=epochs,

       verbose=1,

       validation_data=(X_test, y_test),

       callbacks=[history])

y_test_pred = model.predict(X_test, batch_size=128, verbose=1)

y_label = data.target_names.tolist()

y_test_c = np.argmax(y_test, axis=1).reshape(-1, 1)

y_test_pred_c = np.argmax(y_test_pred, axis=1).reshape(-1, 1)

"""

모델 학습 후 결과를 저장하고 해당 결과를 UI 상에 노출

"""

aistudio_framework.plot_confusion_matrix(y_test_c, y_test_pred_c, target_names=y_label, title='Confusion Matrix')

aistudio_framework.classification_report(y_test_c, y_test_pred_c, target_names=y_label)

aistudio_framework.plot_roc_curve(y_test, y_test_pred, len(y_label), y_label)

'''

학습된 모델 저장

'''

aistudio_framework.save_model(model, model_name='iris-classification')           

from __future__ import print_function

import numpy as np                

import pandas as pd                

"""

AI 포탈과 연계를 위한 기본 객체 생성                

"""

from aicentro.session import Session                

session = Session(verify=False)                

"""

학습 모델 개발 시 프레임워크별 객체 사용                

"""

from aicentro.framework.framework import BaseFramework as Frm                

framework = Frm(session=session)                

"""

모델 코드 작성                

"""

from sklearn.linear_model import LogisticRegression                

from sklearn import datasets                

from sklearn.datasets import load_iris                

from sklearn.preprocessing import LabelEncoder                

from sklearn.model_selection import train_test_split               

data = load_iris()               

X = data.data                

y = data.target                

y_label = data.target_names.tolist()                

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=1)                

logreg = LogisticRegression(C=1e5)                

logreg.fit(X_train, y_train)                

y_test_pred = logreg.predict(X_test)

y_test_c = y_test.reshape(-1, 1)                

y_test_pred_c = y_test_pred.reshape(-1, 1)