AWS Lambda를 통한 Tensorflow 및 Keras 기반 추론 모델 서비스하기

© 2018, Amazon Web Services, Inc. or Its Affiliates. All
rights reserved. ੉ળߧ CFPNJ!OFYPODPLS /&90/,PSFB "84-BNCEBܳాೠ 5FOTPSGMPX,FSBTӝ߈୶ۿݽ؛ࢲ࠺झೞӝ

rights reserved. ੷ח ੉ળߧCFPNJ!OFYPODPLS *OUFMMJHFODF-BCT "OUJBCVTJOH5FBN ؘ੉ఠ࠙ࢳ ਢ ূ૑פয݂ XJUI1:5)0/

rights reserved. য়ט੉ঠӝೡղਊٜ ؀ࣁח٩۞׬,FSBT۽ࡅܰѱ٩۞׬ݽ؛ٜ݅੗ ,FSBT9DFQUJPOݽ؛ӝ߈$MBTTJGJDBUJPOݽ؛ٜ݅ӝ %PDLFS "NB[PO-JOVY-BNCEB1BDLBHF;*1 "84-BNCEB 4 "1*(BUFXBZਢࢲ࠺झ 2O" %FFQ-FBSOJOH!/&90/

rights reserved. য়ט੄ҙ੼ "84࠺ਊѐߊ੗࠙ࢳо੄दр࠺ਊ औѱ঳ח࠙࢑୊ܻ ੺؀੸ੋ੉੼ࠁ׮ח࢚؀੸੉੼

rights reserved. ؀ࣁח٩۞׬   ,FSBT۽ࡅܰѱ٩۞׬ݽ؛ٜ݅੗

rights reserved. חঌѷחؘ ,FSBTחޥоਃ

,FSBT • 5FOTPSGMPX$/5,5IFBOPਤীࢲز੘ೞח  Ҋࣻળ/FVSBM/FU"1*ۄ੉࠳۞ܻ • 5FOTPSGMPXࠗఠղ੢ؽ ߹بࢸ஖೙ਃ9 • /FVSBM/FUਸ੸਷௏٘۽ҳഅоמ
• *NBHF/FUӝ߈ਵ۽1SFUSBJOFEػݽ؛ٜ • ਃ્ীחъച೟णҗҙ۲೧ࢲبਬݺ೧૗

,FSBT۽ࡅܰѱݽ؛ٜ݅ӝ • 4FRVFOUJBMݽ؛ୡ݅ীࢎਊೞӝ 4FRVFOUJBM.PEFMࢤࢿ model = Sequential() model.add(Dense(1, activation='sigmoid', input_dim=100))
from tensorflow.python.keras.models import Sequential from tensorflow.python.keras.layers import Dense import numpy as np 4FRVFOUJBM.PEFM*NQPSU .PEFMDPNQJMF model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy']) 4BNQMFEBUB data = np.random.random((1000, 100)) labels = np.random.randint(2, size=(1000, 1)) .PEFM'JUUJOH model.fit(data, labels, epochs=10, batch_size=32)

*NBHF/FU1SFUSBJOFE.PEFMTJO,FSBT • 9DFQUJPO • 7(( • 7(( • 3FT/FU •
*ODFQUJPOW • *ODFQUJPO3FT/FUW • .PCJMF/FUW

*NBHF/FU1SFUSBJOFE.PEFMTJO,FSBT 5PEBZT$IPJDF • 9DFQUJPO • 7(( • 7(( • 3FT/FU
• *ODFQUJPOW • *ODFQUJPO3FT/FUW • .PCJMF/FUW

*NBHF/FU1SFUSBJOFE.PEFMTJO,FSBT • ࢜ݽ؛ਸJNQPSUೞח҃਋1SFUSBJOFE.PEFMਸ׮਍۽٘ • ೞ૑݅पઁழझథݽ؛ਸࢲࡂೞח҃਋ীח  ݽ؛౵ੌਸ۽ஸী੷੢ೞѢա഑਷T١ਸా೧ࢲࡂ • ইې௏٘৬э੉рױೞѱ.PEFMਸࢎਊоמ from tensorflow.python.keras.applications.xception
import Xception model = Xception(weights='imagenet') ୶о೟णহ੉*NBHF/FUӝ߈ਵ۽о઺஖೟णػݽ؛

rights reserved. 9DFQUJPO৬ೣԋೞח  ݽ؛ઁ੘৘द

9DFQUJPO • ,FSBTઁ੘੗੉੗ݫੋప੉ցੋ'SBOçPJT$IPMMFUоઁ੘ • $//ӝ߈*NBHF/FU1SFUSBJOFEݽ؛ • %FQUIXJTF4FQBSBCMF$POWPMVUJPOبੑ • ੘਷ݽ؛ࢎ੉ૉ .#
• Ҋࢿמ

9DFQUJPOJO,FSBT • ,FSBTীղ੢غযҍ߄۽ࢎਊоמ ੗زਵ۽ݽ؛׮਍۽٘ from tensorflow.python.keras.applications.xception import Xception from tensorflow.python.keras.preprocessing
import image from tensorflow.python.keras.applications.resnet50 import preprocess_input, decode_predictions import numpy as np model = Xception(weights='imagenet') img_path = 'elephant.jpg' img = image.load_img(img_path, target_size=(224, 224)) x = image.img_to_array(img) x = np.expand_dims(x, axis=0) x = preprocess_input(x) preds = model.predict(x) print('Predicted:', decode_predictions(preds, top=3)[0])

9DFQUJPOJO,FSBT • 9DFQUJPO੗୓ח*NBHF/FUਸਤ೧ҳࢿػݽ؛ • ೟णؘ੉ఠܳ֍যળٍݽ؛ਸ೟णदெঠೠ׮ • ؘ੉ఠࣇ੉݆׮ݶ୊਺ࠗఠ࠼ݽ؛ਸ೟णदఃחѪ੉оמ • ೞ૑݅അप੸ਵ۽חજ਷ؘ੉ఠࣇҳೞӝоয۰਑  ݆ਵݶݻୌѐ
૒੽ٜ݅ӝبয۰਑

੸਷ؘ੉ఠ۽֫਷੿ഛبXFJHIUദٙೞӝ • ੸਷ؘ੉ఠծ਷੿ഛب • 9DFQUJPO৬э਷1SFUSBJOFEXFJHIUਸоઉ৬  5SBOTGFS-FBSOJOHਸా೧੸਷ؘ੉ఠ۽ب  ؘ੉ఠࣇীౠചػݽ؛о઺஖ദٙоמ • ؘ੉ఠনਸૐ಩ೞӝ  ੸਷ؘ੉ఠܳઑӘঀ߸ഋೞݴࢎਊ 
IUUQTCMPHLFSBTJPCVJMEJOHQPXFSGVMJNBHFDMBTTJGJDBUJPONPEFMTVTJOHWFSZMJUUMFEBUBIUNM

rights reserved. ஖అҗೖ੗ܳҳ࠙ೞחݽ؛੉੓׮ݶ

9DFQUJPO.PEFM5SBJOJOH৘द • ஖అࢎ૓੢җೖ੗ࢎ૓੢ ೟णਊ • ஖అࢎ૓੢җೖ੗ࢎ૓੢ పझ౟ਊ •
9DFQUJPOݽ؛ӝ߈5SBOTGFS-FBSOJOH from tensorflow.python.keras.applications.xception import Xception model = Xception(include_top=False, weights='imagenet') 'VMM$POOFDUFE -BZFSઁѢ 5SBOTGFS-FBSOJOH

ইې৬э੉ؘ੉ఠо੓׮ݶ 1:5)0/ ೐۽ં౟ಫ؊ USBJOJOHQZ USBJOJOHQJ[[B  ೖ੗ࢎ૓౟ۨ੉׬ࣇ USBJOJOHDIJDLFO  ஖అࢎ૓౟ۨ੉׬ࣇ UFTUQJ[[B  ೖ੗ࢎ૓పझ౟ࣇ
UFTUDIJDLFO  ஖అࢎ૓పझ౟ࣇ

౟ۨ੉׬ೣࣻ def train(train_data_dir, validation_data_dir, model_path): base_model = Xception(weights='imagenet', include_top=False) x
= base_model.output x = GlobalAveragePooling2D()(x) predictions = Dense(nb_classes, activation='softmax')(x) model = Model(base_model.input, predictions) transformation_ratio = .05 train_datagen = ImageDataGenerator(rescale=1. / 255, rotation_range=transformation_ratio, shear_range=transformation_ratio, zoom_range=transformation_ratio, cval=transformation_ratio, horizontal_flip=True, vertical_flip=True) पઁ೟णदఆݽ؛ ౟ۨ੉׬ࣇ੉޷૑ܳ߸ഋ೧ؘ੉ఠ۝ૐ಩ 9DFQUJPO߬੉झ

౟ۨ੉׬ೣࣻ def train(train_data_dir, validation_data_dir, model_path): ...(ࢤۚ)... train_generator = train_datagen.flow_from_directory(train_data_dir, batch_size=32,
class_mode='categorical') validation_generator = validation_datagen.flow_from_directory(validation_data_dir, batch_size=32, class_mode='categorical') model.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy']) top_weights_path = os.path.join(os.path.abspath(model_path), 'top_model_weights.h5') callbacks_list = [ ModelCheckpoint(top_weights_path, monitor='val_acc', verbose=1, save_best_only=True), EarlyStopping(monitor='val_acc', patience=5, verbose=0) ] model.fit_generator(train_generator, samples_per_epoch=train_generator.nb_sample, nb_epoch=nb_epoch / 5, validation_data=validation_generator, nb_val_samples=validation_generator.nb_sample, callbacks=callbacks_list) ૐ಩ؘ੉ఠࢤࢿӝ ݽ؛$PNQJMF ݽ؛5SBJOJOH

౟ۨ੉׬ਸ݃஖Ҋ • ݽ؛౟ۨ੉׬ਸ݃஖ݶKTPO౵ੌҗI౵ੌ۽&YQPSU  ппݽ؛җݽ؛੄о઺஖ • पઁ*OGFSFODFܳ૓೯ೡٸח  ਤীࢲ&YQPSUೠݽ؛җݽ؛о઺஖݅ਸ੍যࢎਊ final_weights_path =
os.path.join(os.path.abspath(model_path), 'model_weights.h5') model.save_weights(final_weights_path) model_json = model.to_json() json_file = open(os.path.join(os.path.abspath(model_path), 'model.json'), 'w') json_file.write(model_json) ݽ؛੷੢ XFJHIU੷੢

ݽ؛5FTU"DDVSBDZஏ੿ೞӝ def inference(trained_model_dir, test_data_dir, results_path): # load json and create
model json_file = open(os.path.join(trained_model_dir, model_name), 'r') loaded_model_json = json_file.read() json_file.close() model = model_from_json(loaded_model_json) model.load_weights(os.path.join(trained_model_dir, model_weights)) # Read Data test_datagen = ImageDataGenerator(rescale=1. / 255) test_generator = test_datagen.flow_from_directory(test_data_dir, batch_size=batch_size, shuffle=False) # Calculate class posteriors probabilities y_probabilities = model.predict_generator(test_generator, val_samples=test_generator.nb_sample) # Calculate class labels y_classes = probas_to_classes(y_probabilities) filenames = [filename.split('/')[1] for filename in test_generator.filenames] ids = [filename.split('.')[0] for filename in filenames] ݽ؛੍ӝ о઺஖੍ӝ

୊਺ࠗఠ૒੽݅ٚݽ؛ਸॳҊरযਃ • ,FSBTоইצ5FOTPSGMPX݅ਵ۽੘ࢿೠݽ؛بࢎਊоמ • 5FOTPSGMPXࡺ݅ইפۄ1Z5PSDIبࢎਊоמ • ೟ण਷(16ജ҃ীࢲ   *OGFSFODFࢲ࠺झח-BNCEBীࢲ

rights reserved. ٩۞׬ݽ؛ਸ  -BNCEBীࢲج۰ࠁҊरযਃ

rights reserved. Ӓր&$ীࢲجܻݶউغաਃ ৵-BNCEBܳॳաਃ

৵-BNCEBܳॳաਃ • -BNCEBחࢲߡܻझইఃఫ୊ܳҳࢿೞחؘࢎਊ • ࢲߡҙܻী؀ೠࠗ׸੉੸਺ • ߹ب੄۽٘ߖ۠ࢲҳࢿহ੉ࢲ࠺झ੄߽۳୊ܻചоמ • ࠗೞо੸׼൤Ѧܻח҃਋ূ૑פয݂੉गо੸਺  ೠߣҳઑܳੜ੟ਵݶূ૑פযب਑হ੉ 
ؘ੉ఠ࠙ࢳооݽ؛ࢲ࠺झয়೑߂ਬ૑оמ

rights reserved. -BNCEBWT&$ ঱ઁ ޖ঺ਸॄঠೞա

-BNCEBWT&$ • زद୊ܻܳрಞೞѱ"84ױীࢲ੸ਊоמ  ௏٘ࣻ੿੉೙ਃೞ૑ঋ਺ • ۈ׮ೣࣻח੹୓ೣࣻী؀೧3FHJPO߹  زदীѐೣ߽ࣻ۳प೯оמ • "844VQQPSU$FOUFSܳా೧زद୊ܻ۝ૐооמ
• $16৬(16ীٮۄݽ؛੄ࣘبର੉о௼૑ঋ਷҃਋  *NBHF/FU*OGFSFODF.PEFM 0   /FVSBM4UZMF5SBOTGFS 9     (16۽بוؘܽ दрت-BNCEBח9

• ੄ઓಁః૑ನೣ.#ղ۽ҳࢿೞחѪ੉оמೠ҃਋  ಁః૚૒റ߄۽ߓನоמ • ݽ؛੄ਊ۝੉੸׼ೠ҃਋ .#_.#   4ীࢲ-BNCEB۽оઉয়חदр-BNCEBप೯࠺ਊ • ೠߣীৈ۞ѐܳزदীप೯ೡ߽۳୊ܻо઺ਃೠ҃਋ 
ೞաप೯दр੉௼ѱ઺ਃೞ૑ঋ਷҃਋ -BNCEBWT&$

-BNCEBWT&$ • 'BSHBUFܳ੉ਊ೧ࢲߡܻझ৬ਬࢎೞѱҳࢿоמ  ইఃఫ୊ҙܻੋ۱੉೙ਃ • ױة&$ܳࢎਊೡ҃਋೦࢚ெઉ੓׮ݶ0O%FNBOE   4-"ࠁ੢੉೙ਃহ׮ݶ4105*OTUBODF۽ҳࢿоמ  ".*ܳ੉ਊೡࣻب੓਺
• (16ੋझఢझܳࢎਊೡࣻ੓਺  ୭ࣗदр׼

• ੄ઓಁః૑ನೣ.#ղ۽ҳࢿೞחѪ੉ࠛоמೠ҃਋  ੄ઓಁః૑ܳଂѐUNQী֍য୊ܻоמೞ૑݅  ࠛಞೞҊ-BNCEBप೯दрਸ૑ো ت • ݽ؛੄ਊ۝੉௾҃਋ ৈ۞ݽ؛ਸࢎਊೞח҃਋  UNQীח୭؀.#݅֍ਸࣻ੓਺ 
7((੄҃਋ݽ؛݅.# • ೞաೞաѤী؀೧ࡅܲ୊ܻо೙ਃ೧(16о೙ਃೠ҃਋  *OQVU%BUBী؀ೠ୊ܻо೙ਃೠ҃਋١ -BNCEBWT&$

rights reserved. 9DFQUJPOݽ؛*OGFSFODFೣࣻ -BNCEBೣࣻ۽ٜ݅ӝ

-BNCEBೣࣻী੹׳غח੿ࠁ • -BNCEBחFWFOU৬DPOUFYUё୓ܳా೧੉߮౟ܳ੹׳ • ਋ܻоࢎਊೞח4'JMFVQMPBEFEחFWFOUё୓۽੹׳

# event ё୓ { 'Records': [ { 'eventVersion': '2.0', 'eventSource':
'aws:s3', 'awsRegion': 'ap-northeast-2', # ߡఉ ܻ੹ 'eventTime': '2017-12-13T03:28:13.528Z', # স۽٘ ৮ܐ दп 'eventName': 'ObjectCreated:Put', 'userIdentity': {'principalId': 'AFK2RA1O3ML1F'}, 'requestParameters': {'sourceIPAddress': '123.24.137.5'}, 'responseElements': { 'x-amz-request-id': '1214K424C14C384D', 'x-amz-id-2': 'BOTBfAoB/gKBbn412ITN4t2psTW499iMRKZDK/CQTsjrkeSSzSdsDUMGabcdnvHeYNtbTDHoHKs=' }, 's3': { 's3SchemaVersion': '1.0', 'configurationId': 'b249eeda-3d48-4319-a7e2-853f964c1a25', 'bucket': { 'name': 'aws-summit-kr-2018', # ߡఉ ੉ܴ 'ownerIdentity': { 'principalId': 'AFK2RA1O3ML1F' }, 'arn': 'arn:aws:s3:::aws-summit-kr-2018' }, 'object': { 'key': 'img/test_img.png', # ߡఉ ղ ౵ੌ੄ ҃۽ 'size': 11733, # ౵ੌ ௼ӝ 'eTag': 'f2d12d123aebda1cc1fk17479207e838', 'sequencer': '125B119E4D7B2A0A48' } } } ] } #VDLFU /BNF 'JMF1BUI JO#VDLFU

-BNCEBীৢܾ)BOEMFSೣٜࣻ݅ӝ • &WFOUੋ੗ীࢲ੿ࠁ঳যয়ӝ • ইې৬э੉FWFOUղ3FDPSETܳా೧੿ࠁ঳ӝ • ੉ઁCVDLFU@OBNF৬GJMF@QBUIܳా೧౵ੌঘࣁझоמ • ౵ੌ౵ੌ0CKFDUܳоઉৢٸח#PUPਸ੉ਊ #
ਪࠗ࠙ ࢤۚ def handler(event, context): bucket_name = event['Records'][0]['s3']['bucket']['name'] file_path = event['Records'][0]['s3']['object']['key']

4ীࢲ౵ੌ׮਍۽٘স۽٘ • #PUPਸ੉ਊ೧౵ੌ׮਍۽٘߂স۽٘ ACCESS_KEY = os.environ.get('ACCESS_KEY') SECRET_KEY = os.environ.get('SECRET_KEY') def
downloadFromS3(strBucket, s3_path, local_path): s3_client = boto3.client('s3', aws_access_key_id=ACCESS_KEY, aws_secret_access_key=SECRET_KEY) s3_client.download_file(strBucket, s3_path, local_path) def uploadToS3(bucket, s3_path, local_path): s3_client = boto3.client('s3', aws_access_key_id=ACCESS_KEY, aws_secret_access_key=SECRET_KEY) s3_client.upload_file(local_path, bucket, s3_path) -BNCEBী޷ܻജ҃߸ࣻ۽"84ఃद௼݁١۾

• খࢲ݅ٚIBOEMFSೣࣻղ੉޷૑׮਍۽٘୶о • AUNQAಫ؊ղ੷੢ .ઁೠ   ഑਷ৡݫݽܻ'JMF0CKFDU۽ҳࢿ def handler(event, context):
bucket_name = event['Records'][0]['s3']['bucket']['name'] file_path = event['Records'][0]['s3']['object']['key'] file_name = file_path.split('/')[-1] downloadFromS3(bucket_name, file_path, '/tmp/'+file_name) 4ীࢲ౵ੌ׮਍۽٘স۽٘ -BNCEBীࢲ'JMFXSJUFח  AUNQAղীࢲ݅оמ ׮ܲҔীॳݶ*0&SSPSߊࢤ

9DFQUJPOݽ؛׮਍۽٘ • 4ߡఉীࢲݽ؛XFJHIU౵ੌ I ׮਍۽٘ def handler(event, context): bucket_name =
event['Records'][0]['s3']['bucket']['name'] file_path = event['Records'][0]['s3']['object']['key'] file_name = file_path.split('/')[-1] downloadFromS3(bucket_name, file_path, '/tmp/'+file_name) downloadFromS3( 'aws-summit-kr-2018', 'xception_weights_tf_dim_ordering_tf_kernels.h5', '/tmp/.keras/xception_weights_tf_dim_ordering_tf_kernels.h5' ) ,FSBTח_LFSBT৬UNQLFSBTفҔীࢲ౵ੌਸ଺ইࢎਊೞ૑݅ -BNCEBীࢲח_LFSBTীॳӝӂೠ੉হਵ޲۽UNQLFSBTܳࢎਊ

9DFQUJPO1SFEJDUೣࣻ • AQSFEJDUAۄח੉ܴਵ۽9DFQUJPOݽ؛ࢎਊೞӝ from tensorflow.python.keras.applications.xception import Xception from tensorflow.python.keras.preprocessing import
image from tensorflow.python.keras.applications.resnet50 import preprocess_input, decode_predictions import numpy as np def predict(img_path): model = Xception(weights='imagenet') img = image.load_img(img_path, target_size=(224, 224)) x = image.img_to_array(img) x = np.expand_dims(x, axis=0) x = preprocess_input(x) preds = model.predict(x) return decode_predictions(preds, top=3)[0] ੉޷૑҃۽ܳੋ੗۽߉ই1SFEJDUܳ݃஘Ѿҗчਸ߈ജ ױࣽೠѾҗч3FUVSO؀न %ZOBNP%#١ী੸੤੉޷૑࠙ܨറಫ؊߹4স۽٘

׮दೠߣ-BNCEBݫੋೣࣻܳࣻ੿ೞݶ • 4&WFOU%PXOMPBE*NBHF.PEFM1SFEJDU def handler(event, context): bucket_name = event['Records'][0]['s3']['bucket']['name'] file_path
= event['Records'][0]['s3']['object']['key'] file_name = file_path.split('/')[-1] downloadFromS3(bucket_name, file_path, '/tmp/'+file_name) downloadFromS3( 'aws-summit-kr-2018', 'xception_weights_tf_dim_ordering_tf_kernels.h5', '/tmp/.keras/xception_weights_tf_dim_ordering_tf_kernels.h5' ) result = predict('/tmp/'+file_name) return result

rights reserved. -BNCEBস۽٘ਊ QBDL[JQٜ݅ӝ ೙ਃೠ1ZUIPO%FQFOEFODZٜ੉ ݽفݽৈ੓ח঑୷౵ੌٜ݅ӝ

rights reserved. ੄ઓಁః૑о݆ইਃ

rights reserved. $঱য҅ৌ੄ઓۄ੉࠳۞ܻ 1BOEBT/VNQZIQZ

rights reserved. 04߹۽׮ܲXIM౵ੌ

rights reserved. ࠽٘

rights reserved. "84-BNCEB04 "NB[PO-JOVY

-BNCEBPO"NB[PO-JOVY • ".*BN[OBNJIWNY@HQ • -JOVYழօBN[OY@ • CJU߄੉ցܻ݅૑ਗ • 1ZUIPO਷ࢎਊоמ  #PUP
CPUPDPSF ࢸ஖غয੓਺ • $FOU04୊ۢ:VNࢎਊоמ ୭न੿ࠁ  IUUQTEPDTBXTBNB[PODPNMBNCEBMBUFTUEHDVSSFOUTVQQPSUFEWFSTJPOTIUNM

"84-BNCEB1BDLBHJOH • -BNCEBী౵੉ॆ౵ੌਸৢܻӝਤ೧ࢲח  ೠ[JQ౵ੌী੄ઓࢿಁః૑ܳݽفޘযࢲৢ۰ঠೣ • $PEF%FQFOEFODZਊ۝ઁೠ  "84$POTPMFীࢲস۽٘оמೠ[JQ୭؀௼ӝ.#  ؊௾ਊ۝਷4ਸѢ୛স۽٘оמ  
঑୷೧ઁೠ࢚క.#੉ղ 9DFQUJPOݽ؛بড.# ݽ؛ਸэ੉ಁః૚ೞחѪ਷য۰ਕਃ

• 5FOTPSGMPX /VNQZ 1BOEBT١$MBOH੄ઓ੸ੋಁః૑ח  04ীݏ୾࠽٘੘সਸ૓೯೧ঠೣ • *TTVFѐߊജ҃਷.BD048JOEPXT഑਷6CVOUVੋؘ  "84-BNCEB਷"NB[PO-JOVYਤীࢲҳز઺ • ٮۄࢲ"NB[PO-JOVYਤীࢲ࠽٘೙ਃ
• &$ܳࢎਊ೧ழझథ࠽٘୶о࠺ਊߊࢤ • %PDLFSܳా೧۽ஸী"NB[PO-JOVYജ҃ҳ୷റ࠽٘ "84-BNCEB1BDLBHJOH

rights reserved. بழо೙ਃೞ׮

بழח৵೙ਃೠоਃ • بழஶప੉ցӝ߈੄য়೑ࣗझо࢚ച೒ۖಬ • ӝઓ7.੄ࢿמ੷ೞহ੉ழօਸҕਬ য়ߡ ೻٘ ੸਺

• 7.ਸ੉ਊ೧"NB[PO-JOVYܳࢎਊ೧بغ૑݅  ࠽٘ࣘبо࢚؀੸ਵ۽بழী࠺೧וܿ  दझమܻࣗझܳ׮۝ࢎਊ • بழীחब૑য"NB[PO-JOVY੉޷૑ب੓׮  ݺ۸যೠ઴۽"NB[PO-JOVYஶప੉ցܳڸ਎ࣻ੓׮ • -BNCEBীৢܾ[JQಂਸٜ݅ٸبழܳࢎਊಞܻ بழח৵೙ਃೠоਃ

rights reserved. "NB[PO-JOVY

"NB[PO-JOVYPO%PDLFS • AEPDLFSAۄחݺ۸যܳࢎਊೡࣻ੓ب۾  %PDLFS$PNNVOJUZ&EJUJPOਸࢸ஖೤פ׮ • ইېݺ۸য۽"NB[PO-JOVY੉޷૑ӝ߈  بழஶప੉ցܳप೯ೡࣻ੓णפ׮ • ਤݺ۸যܳࢎਊೞݶ࢜بழஶప੉ցоڸਕ૑ݴ  
അ੤ಫ؊ܳبழஶప੉ցղ੄APVUQVUTAীোѾ೤פ׮ docker run -v $(pwd):/outputs --name lambdapack -d amazonlinux:latest tail -f /dev/null

بழղীࢲप೯ೡ࠽٘4IFMM4DSJQUٜ݅ӝ • :VNਵ۽ಁః૑সؘ੉౟౵੉ॆࢸ஖ • ౵੉ॆо࢚ജٜ҃݅ҊSFRVJSFNFOUTࢸ஖ • о࢚ജ҃TJUFQBDLBHFTղݽٚ౵ੌ੉ز  ױ UFTU౵ੌҗDBDIF౵ੌ਷ઁ৻
• ۽ஸী੓ח -BNCEBীৢۄт1ZUIPO౵ੌࠂࢎ • 1ZUIPO౵ੌҗಁః૑੹୓঑୷  ઱੄ಫ؊оইצ౵ੌٜ੹୓঑୷

%PDLFSীZVNਵ۽೙ਃಁః૑ࢸ஖ dev_install() { yum -y update yum -y upgrade yum
install -y \ wget \ gcc \ gcc-c++ \ cmake \ python36-devel \ python36-virtualenv \ python36-pip \ findutils \ zlib-devel \ zip \ unzip \ blas-devel lapack-devel atlas-devel } ӝࠄಁః૑୭नসؘ੉౟ 1ZUIPOࢸ஖ ӝఋಁః૑ࢸ஖

%PDLFSղWJSUVBMFOW۽೙ਃಁః૑݅ࢸ஖ mkvirtualenv() { cd /home/ rm -rf env python3 -m
virtualenv env --python=python3 source env/bin/activate } pip_install() { source /home/env/bin/activate pip install -U pip wheel pip install --use-wheel tensorflow==1.7.0 --no-deps pip install protobuf html5lib bleach --no-deps pip install --use-wheel pillow==4.0.0 pip install h5py } FOWо࢚ജ҃ࢤࢿ ೙ਃ੄ઓࢿಁః૑ࢸ஖ OPEFQT੄ઓಁః૑ࢸ஖9 ೧׼ಁః૑݅ࢸ஖

EFQFOEFODZ JOEFYQZQBDL[JQ gather_pack() { cd /home/ && rm -rf pack
&& mkdir pack && cd pack cp -R /home/env/lib/python3.6/site-packages/* . cp -R /home/env/lib64/python3.6/site-packages/* . cp /outputs/index.py /home/pack/index.py find . -type d -name "test" -exec rm -rf {} + find -name "*.so" | xargs strip find -name "*.so.*" | xargs strip rm -r pip && rm -r pip-* && rm -r wheel && rm -r wheel-* find . | grep -E "(__pycache__|\.pyc$)" | xargs rm -rf echo "stripped size $(du -sh /home/pack | cut -f1)" zip -FS -r1 /outputs/pack.zip * > /dev/null echo "compressed size $(du -sh /outputs/pack.zip | cut -f1)" } ঑୷ೡಫ؊ࢤࢿ ౵੉ॆۄ੉࠳۞ܻࠂࢎ पઁز੘ೞח JOEFYQZ 4IBSFE0CKFDU TP TUSJQ QZDBDIFઁѢ QBDL[JQਵ۽঑୷

"84௑ࣛীࢲ-BNCEBೣٜࣻ݅ӝ • %FQFOEFODZоݽفࢸ஖ػQBDL[JQ౵ੌਸ4ীস۽٘ • "NB[PO4ীࢲ౵ੌস۽٘২࣌ਵ۽QBDL[JQস۽٘ • 3VOUJNF1ZUIPO • )BOEMFS౵੉ॆ౵ੌ੉ܴೣࣻ੉ܴ •
ݫݽܻ(੉࢚ઁೠदрୡ੉࢚ • ജ҃߸ࣻࢸ੿റ੷੢

• QBDL[JQ౵ੌ੉.#੉ೞੋ҃਋  ۽ஸBXTDMJജ҃ীࢲ߄۽ೣࣻসؘ੉౟оמ • QBDL[JQ౵ੌ੉.#ୡҗೞח҃਋  4ী౵ੌস۽٘੉റ4҃۽ #VDLFU ,FZ ా೧সؘ੉౟ •
౵ੌস۽٘৬ೣࣻসؘ੉౟ݽف#PUP۽૓೯ 46QMPBE-BNCEB'VODUJPOVQEBUF

46QMPBE-BNCEB'VODUJPOVQEBUFX#PUP import boto3 def upload_to_s3(bucket, s3_path, local_path): client = boto3.client('s3',
aws_access_key_id=ACCESS_KEY,aws_secret_access_key=SECRET_KEY) client.upload_file(local_path, bucket, s3_path) def update_lambda(function_name, bucket, s3_path): client = boto3.client('lambda', aws_access_key_id=ACCESS_KEY,aws_secret_access_key=SECRET_KEY) client.update_function_code( FunctionName=function_name, S3Bucket=bucket, S3Key=s3_path, ) uploadToS3('ߡఉݺ', 'ߡఉղ҃۽/pack.zip', './pack.zip') update_lambda('ۈ׮ೣࣻ੉ܴ', 'ߡఉݺ', 'ߡఉղ҃۽/pack.zip') QBDL[JQস۽٘ MBNCEBೣࣻসؘ੉౟

rights reserved. %FQFOEFODZܳݽف೤ଢ଼؊פ  .#оֈযߡ۷যਃ

.#੄ೠ҅ܳֈয • <઱੄>୶ୌ9 • ೞ૑݅աחԙ-BNCEBীࢲجܻѷয  ഑਷ৈӝࢲޛ۞զࣻחহ׮ۄחࢤп੉दۄݶ  • 1BDL[JQفѐ۽ଂѐӝ %FQFOEFODZ݅੓חQBDL[JQ
• -BNCEBೣࣻ౵ੌઁੌখীଂђEFQFOEFODZ[JQ౵ੌ߉ӝ • UNQղ঑୷೧ઁ • TZTQBUIীUNQಫ؊١۾

%FQFOEFODZଂѐӝ def download_s3_object(strBucket, s3_path): import io file_obj = io.BytesIO() s3_client
= boto3.client( 's3', aws_access_key_id=ACCESS_KEY, aws_secret_access_key=SECRET_KEY ) s3_client.download_fileobj(strBucket, s3_path, file_obj) return file_obj pack2 = download_s3_object('bucket-name', 'pack2.zip') import zipfile zip_ref = zipfile.ZipFile(pack2) zip_ref.extractall('/tmp') zip_ref.close() import sys sys.path.append("/tmp") ౵ੌਸ׮਍߉૑݅ ৡݫݽܻ੄#ZUFT*0ী੷੢ UNQಫ؊ী %FQFOEFODZ঑୷೧ઁ пઙۄ੉࠳۞ܻJNQPSU੹ JNQPSU1"5)୶о

• ٩۞׬ݽ؛җ%FQFOEFODZ౵ੌ਷য٣ࢲоઉয়ա 4 • 4ীࢲоઉয়ݶ੹࣠࠺ਊ਷হաਃ   э਷3FHJPOղ੹࣠਷੹࣠࠺ਊ੉ߊࢤೞ૑ঋणפ׮ • 4ীࢲоઉয়חؘदр੉ٜ૑ঋաਃ  
-BNCEBदр੉ઑӘ؊য়ېѦܻ૑݅   ௾ਊ۝੄ݽ؛౵ੌਸࢎਊೡࣻ੓׮ח੼੉੓णפ׮ • ਊ۝ઁೠ਷হաਃ   पઁ౵ੌ۽ॵ׮ݶ୭؀.# ৡݫݽܻ۽୭؀(੉ղ पઁ۽ࢎਊೞӝীޙઁ੓૑ঋաਃ

rights reserved. -BNCEB 4 "1*(BUFXBZ  8FC4FSWJDF

rights reserved. -BNCEB5SJHHFSJOH &WFOU

-BNCEB5SJHHFSJOH • -BNCEBח੉߮౟ী੄೧ഥप೯غח  4UBUFMFTT'VODUJPO • ٮۄࢲ&WFOUী੿ࠁܳ׸ই੹׳ • ׮নೠ&WFOU5SJHHFSо੓૑݅   ੉ߣীח4VQMPBEFEࢎਊ
• ౠ੿ߡఉղౠ੿1SFGJY ಫ؊ ղ  ࢜۽਍౵ੌ੉ࢤӝח҃਋ -BNCEBܳप೯ೞ੗

-BNCEB5SJHHFSJOH ࢎ૓ਸ ৢܻݶ 4ীস۽٘ ࢜੉߮౟ -BNCEB ౟ܻѢ -BNCEBೣࣻоز੘ೞҊ Ѿҗܳ߈ജ //
.PEFM *OGFSFODF

rights reserved. ঒ Ӓؘ۠4ী ੉޷૑ܳযڌѱৢܻ૑

rights reserved. 41045 "1*(BUFXBZ पઁ স۽٘ স۽٘੹ ӂೠࠗৈ

41045 "65)X"1*(BUFXBZ • 4ী౵ੌਸৢܾٸ4҃۽ী1045)551ਃ୒ਸా೧  ౵ੌਸҍ߄۽ৢܻחѪ੉оמ • ٮۄࢲ'SPOUFOEਢಕ੉૑ীࢲױࣽ౵ੌস۽٘ಬਵ۽ب  4ী౵ੌਸস۽٘ೞחѪ੉оמ • ೞ૑݅ߡఉী1VCMJD8SJUFӂೠ੉ࠗৈغয੓૑ঋਵݶ 
ਤ৬э੉ࢎਊೞӝয۰਑ • ٮۄࢲӂೠࠗৈܳਤೠ-BNCEBೣࣻ "1*(BUFXBZ۽  ੋૐ੹ਊ"1*ࢤࢿ

41045 "65)X"1*(BUFXBZ4JOHMF 46QMPBE 4JHOJOH ੉޷૑౵ੌ স۽٘ ੉޷૑оࢤ҂׮ ٩۞׬ਸجܻ੗ ղࢤпীח  ੉ࢎ૓਷஖అ੉׮

41045 "65)X"1*(BUFXBZ.VMUJ [JQস۽٘ ঑୷ਸಽযࠁ੗ ৡݫݽܻ ୭؀( ੉޷૑ࢤӣ ࠺زӝ߽۳୊ܻ ٩۞׬ਸجܻ੗
ݽٚ੉޷૑о ୊ܻغݶ [JQ೧ࠁղ઱੗

rights reserved. ױࣽೞѱ-BNCEBജ҃ীࢲ प೯݅೧ࠁҊर׮ݶ

rights reserved. %PDLFS-BNCEB

rights reserved. docker run --rm -v "$PWD":/var/task lambci/lambda:python3.6 my_module.my_handler ױࣽ൤ೣࣻܳप೯ೞ۰ݶ &WFOUੋ੗ܳ੹׳ೞ۰ݶ docker run --rm -v "$PWD":/var/task lambci/lambda:python3.6 my_module.my_handler '{"some": "event"}'

rights reserved. %FFQ-FBSOJOH!

rights reserved. ѱ੐ഥࢎীࢲ٩۞׬

$MJFOU -PHT 4FSWFS -PHT 0UIFS  -PHT ೞܖ 5#

rights reserved. ࢎۈ੄׀ਸ؀न೧ࠁ੗

rights reserved. '14೨੟ӝ

rights reserved. ࢎۈ੉ࠁ૑ޅೠ1BUUFSOਸ଺੗

rights reserved. $੉਄੉ইצ઴ইפ

rights reserved. ױࣽೠܙ߬੉झࠗఠ ࠂ੟ೠ٩۞׬ө૑

rights reserved. ਋ܻэ੉ੌ೧ਃ 8FSF)JSJOH

AWS Lambda를 통한 Tensorflow 및 Keras 기반 추론 모델 서비스하기

AWS Lambda를 통한 Tensorflow 및 Keras 기반 추론 모델 서비스하기

More Decks by Beomi

Other Decks in Programming

Featured

Transcript