 在 TensorFlow.org 上檢視
|
 在 Google Colab 中執行
|
 在 GitHub 上檢視原始碼
|
 下載筆記本
|
總覽
本教學課程著重於將來自 Kafka 叢集的串流資料導入 tf.data.Dataset，然後將其與 tf.keras 結合使用，以進行訓練和推論。
Kafka 主要是一個分散式事件串流平台，可在資料管線之間提供可擴充且容錯的串流資料。它是許多主要企業的基本技術組件，在這些企業中，任務關鍵型資料傳遞是主要要求。
設定
安裝所需的 tensorflow-io 和 kafka 套件
pip install tensorflow-iopip install kafka-python
匯入套件
import os
from datetime import datetime
import time
import threading
import json
from kafka import KafkaProducer
from kafka.errors import KafkaError
from sklearn.model_selection import train_test_split
import pandas as pd
import tensorflow as tf
import tensorflow_io as tfio
驗證 tf 和 tfio 匯入
print("tensorflow-io version: {}".format(tfio.__version__))
print("tensorflow version: {}".format(tf.__version__))
tensorflow-io version: 0.23.1 tensorflow version: 2.8.0-rc0
下載並設定 Kafka 和 Zookeeper 執行個體
為了示範目的，以下執行個體在本機設定
- Kafka (Brokers: 127.0.0.1:9092)
- Zookeeper (Node: 127.0.0.1:2181)
curl -sSOL https://downloads.apache.org/kafka/2.7.2/kafka_2.13-2.7.2.tgztar -xzf kafka_2.13-2.7.2.tgz
使用預設組態 (由 Apache Kafka 提供) 來啟動執行個體。
./kafka_2.13-2.7.2/bin/zookeeper-server-start.sh -daemon ./kafka_2.13-2.7.2/config/zookeeper.properties./kafka_2.13-2.7.2/bin/kafka-server-start.sh -daemon ./kafka_2.13-2.7.2/config/server.propertiesecho "Waiting for 10 secs until kafka and zookeeper services are up and running"sleep 10
Waiting for 10 secs until kafka and zookeeper services are up and running
執行個體以守護程序啟動後，在程序清單中 grep 搜尋 kafka。這兩個 java 程序對應於 zookeeper 和 kafka 執行個體。
ps -ef | grep kafka
kbuilder 27856 20044 4 20:28 ? 00:00:00 python /tmpfs/src/gfile/executor.py --input_notebook=/tmpfs/src/temp/docs/tutorials/kafka.ipynb --timeout=15000 kbuilder 28271 1 16 20:28 ? 00:00:01 java -Xmx512M -Xms512M -server -XX:+UseG1GC -XX:MaxGCPauseMillis=20 -XX:InitiatingHeapOccupancyPercent=35 -XX:+ExplicitGCInvokesConcurrent -XX:MaxInlineLevel=15 -Djava.awt.headless=true -Xlog:gc*:file=/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../logs/zookeeper-gc.log:time,tags:filecount=10,filesize=100M -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.authenticate=false -Dcom.sun.management.jmxremote.ssl=false -Dkafka.logs.dir=/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../logs -Dlog4j.configuration=file:./kafka_2.13-2.7.2/bin/../config/log4j.properties -cp /tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/activation-1.1.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/aopalliance-repackaged-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/argparse4j-0.7.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/audience-annotations-0.5.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/commons-cli-1.4.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/commons-lang3-3.8.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-api-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-basic-auth-extension-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-file-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-json-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-mirror-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-mirror-client-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-runtime-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-transforms-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/hk2-api-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/hk2-locator-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/hk2-utils-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-annotations-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-core-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-databind-2.10.5.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-dataformat-csv-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-datatype-jdk8-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-jaxrs-base-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-jaxrs-json-provider-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-module-jaxb-annotations-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-module-paranamer-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-module-scala_2.13-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.activation-api-1.2.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.annotation-api-1.3.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.inject-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.validation-api-2.0.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.ws.rs-api-2.1.6.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.xml.bind-api-2.3.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javassist-3.25.0-GA.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javassist-3.26.0-GA.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javax.servlet-api-3.1.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javax.ws.rs-api-2.1.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jaxb-api-2.3.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-client-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-common-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-container-servlet-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-container-servlet-core-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-hk2-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-server-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-client-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-continuation-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-http-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-io-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-security-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-server-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-servlet-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-servlets-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-util-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-util-ajax-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jopt-simple-5.0.4.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-clients-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-log4j-appender-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-raft-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-examples-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-scala_2.13-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-test-utils-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-tools-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka_2.13-2.7.2-sources.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka_2.13-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/log4j-1.2.17.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/lz4-java-1.7.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/maven-artifact-3.8.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/metrics-core-2.2.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-buffer-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-codec-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-common-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-handler-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-resolver-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-transport-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-transport-native-epoll-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-transport-native-unix-common-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/osgi-resource-locator-1.0.3.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/paranamer-2.8.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/plexus-utils-3.2.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/reflections-0.9.12.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/rocksdbjni-5.18.4.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-collection-compat_2.13-2.2.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-java8-compat_2.13-0.9.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-library-2.13.3.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-logging_2.13-3.9.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-reflect-2.13.3.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/slf4j-api-1.7.30.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/slf4j-log4j12-1.7.30.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/snappy-java-1.1.7.7.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/zookeeper-3.5.9.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/zookeeper-jute-3.5.9.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/zstd-jni-1.4.5-6.jar org.apache.zookeeper.server.quorum.QuorumPeerMain ./kafka_2.13-2.7.2/config/zookeeper.properties kbuilder 28635 1 57 20:28 ? 00:00:05 java -Xmx1G -Xms1G -server -XX:+UseG1GC -XX:MaxGCPauseMillis=20 -XX:InitiatingHeapOccupancyPercent=35 -XX:+ExplicitGCInvokesConcurrent -XX:MaxInlineLevel=15 -Djava.awt.headless=true -Xlog:gc*:file=/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../logs/kafkaServer-gc.log:time,tags:filecount=10,filesize=100M -Dcom.sun.management.jmxremote -Dcom.sun.management.jmxremote.authenticate=false -Dcom.sun.management.jmxremote.ssl=false -Dkafka.logs.dir=/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../logs -Dlog4j.configuration=file:./kafka_2.13-2.7.2/bin/../config/log4j.properties -cp /tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/activation-1.1.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/aopalliance-repackaged-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/argparse4j-0.7.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/audience-annotations-0.5.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/commons-cli-1.4.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/commons-lang3-3.8.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-api-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-basic-auth-extension-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-file-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-json-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-mirror-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-mirror-client-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-runtime-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/connect-transforms-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/hk2-api-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/hk2-locator-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/hk2-utils-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-annotations-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-core-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-databind-2.10.5.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-dataformat-csv-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-datatype-jdk8-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-jaxrs-base-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-jaxrs-json-provider-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-module-jaxb-annotations-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-module-paranamer-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jackson-module-scala_2.13-2.10.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.activation-api-1.2.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.annotation-api-1.3.5.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.inject-2.6.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.validation-api-2.0.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.ws.rs-api-2.1.6.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jakarta.xml.bind-api-2.3.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javassist-3.25.0-GA.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javassist-3.26.0-GA.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javax.servlet-api-3.1.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/javax.ws.rs-api-2.1.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jaxb-api-2.3.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-client-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-common-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-container-servlet-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-container-servlet-core-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-hk2-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jersey-server-2.34.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-client-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-continuation-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-http-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-io-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-security-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-server-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-servlet-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-servlets-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-util-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jetty-util-ajax-9.4.43.v20210629.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/jopt-simple-5.0.4.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-clients-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-log4j-appender-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-raft-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-examples-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-scala_2.13-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-streams-test-utils-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka-tools-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka_2.13-2.7.2-sources.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/kafka_2.13-2.7.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/log4j-1.2.17.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/lz4-java-1.7.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/maven-artifact-3.8.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/metrics-core-2.2.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-buffer-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-codec-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-common-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-handler-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-resolver-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-transport-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-transport-native-epoll-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/netty-transport-native-unix-common-4.1.59.Final.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/osgi-resource-locator-1.0.3.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/paranamer-2.8.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/plexus-utils-3.2.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/reflections-0.9.12.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/rocksdbjni-5.18.4.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-collection-compat_2.13-2.2.0.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-java8-compat_2.13-0.9.1.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-library-2.13.3.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-logging_2.13-3.9.2.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/scala-reflect-2.13.3.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/slf4j-api-1.7.30.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/slf4j-log4j12-1.7.30.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/snappy-java-1.1.7.7.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/zookeeper-3.5.9.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/zookeeper-jute-3.5.9.jar:/tmpfs/src/temp/docs/tutorials/kafka_2.13-2.7.2/bin/../libs/zstd-jni-1.4.5-6.jar kafka.Kafka ./kafka_2.13-2.7.2/config/server.properties kbuilder 28821 27860 0 20:28 pts/0 00:00:00 /bin/bash -c ps -ef | grep kafka kbuilder 28823 28821 0 20:28 pts/0 00:00:00 grep kafka
使用以下規格建立 kafka 主題
- susy-train: partitions=1, replication-factor=1
- susy-test: partitions=2, replication-factor=1
./kafka_2.13-2.7.2/bin/kafka-topics.sh --create --bootstrap-server 127.0.0.1:9092 --replication-factor 1 --partitions 1 --topic susy-train./kafka_2.13-2.7.2/bin/kafka-topics.sh --create --bootstrap-server 127.0.0.1:9092 --replication-factor 1 --partitions 2 --topic susy-test
Created topic susy-train. Created topic susy-test.
描述主題以取得組態詳細資訊
./kafka_2.13-2.7.2/bin/kafka-topics.sh --describe --bootstrap-server 127.0.0.1:9092 --topic susy-train./kafka_2.13-2.7.2/bin/kafka-topics.sh --describe --bootstrap-server 127.0.0.1:9092 --topic susy-test
Topic: susy-train PartitionCount: 1 ReplicationFactor: 1 Configs: segment.bytes=1073741824
Topic: susy-train Partition: 0 Leader: 0 Replicas: 0 Isr: 0
Topic: susy-test PartitionCount: 2 ReplicationFactor: 1 Configs: segment.bytes=1073741824
Topic: susy-test Partition: 0 Leader: 0 Replicas: 0 Isr: 0
Topic: susy-test Partition: 1 Leader: 0 Replicas: 0 Isr: 0
複寫因數 1 表示資料未被複寫。這是因為我們的 kafka 設定中只有一個代理程式。在生產系統中，啟動伺服器的數量可能在 100 個節點的範圍內。這就是使用複寫實現容錯的地方。
請參閱 文件 以取得更多詳細資訊。
SUSY 資料集
Kafka 作為事件串流平台，可讓來自各種來源的資料寫入其中。例如
- 網路流量日誌
- 天文測量
- IoT 感測器資料
- 產品評論等等。
為了本教學課程的目的，讓我們下載 SUSY 資料集，並手動將資料饋送到 kafka 中。此分類問題的目標是區分產生超對稱粒子的訊號程序和不產生超對稱粒子的背景程序。
curl -sSOL https://archive.ics.uci.edu/ml/machine-learning-databases/00279/SUSY.csv.gz
探索資料集
第一欄是類別標籤 (1 代表訊號，0 代表背景)，後接 18 個特徵 (8 個低階特徵，然後是 10 個高階特徵)。前 8 個特徵是粒子偵測器在加速器中測量的運動學屬性。最後 10 個特徵是前 8 個特徵的函數。這些是物理學家推導出的高階特徵，有助於區分這兩個類別。
COLUMNS = [
# labels
'class',
# low-level features
'lepton_1_pT',
'lepton_1_eta',
'lepton_1_phi',
'lepton_2_pT',
'lepton_2_eta',
'lepton_2_phi',
'missing_energy_magnitude',
'missing_energy_phi',
# high-level derived features
'MET_rel',
'axial_MET',
'M_R',
'M_TR_2',
'R',
'MT2',
'S_R',
'M_Delta_R',
'dPhi_r_b',
'cos(theta_r1)'
]
整個資料集包含 500 萬列。但是，為了本教學課程的目的，讓我們只考慮資料集的一小部分 (100,000 列)，以便減少花在移動資料上的時間，而將更多時間用於理解 API 的功能。
susy_iterator = pd.read_csv('SUSY.csv.gz', header=None, names=COLUMNS, chunksize=100000)
susy_df = next(susy_iterator)
susy_df.head()
# Number of datapoints and columns
len(susy_df), len(susy_df.columns)
(100000, 19)
# Number of datapoints belonging to each class (0: background noise, 1: signal)
len(susy_df[susy_df["class"]==0]), len(susy_df[susy_df["class"]==1])
(54025, 45975)
分割資料集
train_df, test_df = train_test_split(susy_df, test_size=0.4, shuffle=True)
print("Number of training samples: ",len(train_df))
print("Number of testing sample: ",len(test_df))
x_train_df = train_df.drop(["class"], axis=1)
y_train_df = train_df["class"]
x_test_df = test_df.drop(["class"], axis=1)
y_test_df = test_df["class"]
# The labels are set as the kafka message keys so as to store data
# in multiple-partitions. Thus, enabling efficient data retrieval
# using the consumer groups.
x_train = list(filter(None, x_train_df.to_csv(index=False).split("\n")[1:]))
y_train = list(filter(None, y_train_df.to_csv(index=False).split("\n")[1:]))
x_test = list(filter(None, x_test_df.to_csv(index=False).split("\n")[1:]))
y_test = list(filter(None, y_test_df.to_csv(index=False).split("\n")[1:]))
Number of training samples: 60000 Number of testing sample: 40000
NUM_COLUMNS = len(x_train_df.columns)
len(x_train), len(y_train), len(x_test), len(y_test)
(60000, 60000, 40000, 40000)
將訓練和測試資料儲存在 kafka 中
將資料儲存在 kafka 中模擬了用於訓練和推論目的的持續遠端資料擷取環境。
def error_callback(exc):
raise Exception('Error while sendig data to kafka: {0}'.format(str(exc)))
def write_to_kafka(topic_name, items):
count=0
producer = KafkaProducer(bootstrap_servers=['127.0.0.1:9092'])
for message, key in items:
producer.send(topic_name, key=key.encode('utf-8'), value=message.encode('utf-8')).add_errback(error_callback)
count+=1
producer.flush()
print("Wrote {0} messages into topic: {1}".format(count, topic_name))
write_to_kafka("susy-train", zip(x_train, y_train))
write_to_kafka("susy-test", zip(x_test, y_test))
Wrote 60000 messages into topic: susy-train Wrote 40000 messages into topic: susy-test
定義 tfio 訓練資料集
IODataset 類別用於將資料從 kafka 串流到 tensorflow 中。此類別繼承自 tf.data.Dataset，因此具有 tf.data.Dataset 的所有實用功能。
def decode_kafka_item(item):
message = tf.io.decode_csv(item.message, [[0.0] for i in range(NUM_COLUMNS)])
key = tf.strings.to_number(item.key)
return (message, key)
BATCH_SIZE=64
SHUFFLE_BUFFER_SIZE=64
train_ds = tfio.IODataset.from_kafka('susy-train', partition=0, offset=0)
train_ds = train_ds.shuffle(buffer_size=SHUFFLE_BUFFER_SIZE)
train_ds = train_ds.map(decode_kafka_item)
train_ds = train_ds.batch(BATCH_SIZE)
2022-01-07 20:29:21.602817: E tensorflow/stream_executor/cuda/cuda_driver.cc:271] failed call to cuInit: CUDA_ERROR_NO_DEVICE: no CUDA-capable device is detected
建構和訓練模型
# Set the parameters
OPTIMIZER="adam"
LOSS=tf.keras.losses.BinaryCrossentropy(from_logits=True)
METRICS=['accuracy']
EPOCHS=10
# design/build the model
model = tf.keras.Sequential([
tf.keras.layers.Input(shape=(NUM_COLUMNS,)),
tf.keras.layers.Dense(128, activation='relu'),
tf.keras.layers.Dropout(0.2),
tf.keras.layers.Dense(256, activation='relu'),
tf.keras.layers.Dropout(0.4),
tf.keras.layers.Dense(128, activation='relu'),
tf.keras.layers.Dropout(0.4),
tf.keras.layers.Dense(1, activation='sigmoid')
])
print(model.summary())
Model: "sequential"
_________________________________________________________________
Layer (type) Output Shape Param #
=================================================================
dense (Dense) (None, 128) 2432
dropout (Dropout) (None, 128) 0
dense_1 (Dense) (None, 256) 33024
dropout_1 (Dropout) (None, 256) 0
dense_2 (Dense) (None, 128) 32896
dropout_2 (Dropout) (None, 128) 0
dense_3 (Dense) (None, 1) 129
=================================================================
Total params: 68,481
Trainable params: 68,481
Non-trainable params: 0
_________________________________________________________________
None
# compile the model
model.compile(optimizer=OPTIMIZER, loss=LOSS, metrics=METRICS)
# fit the model
model.fit(train_ds, epochs=EPOCHS)
Epoch 1/10 /tmpfs/src/tf_docs_env/lib/python3.7/site-packages/tensorflow/python/util/dispatch.py:1082: UserWarning: "`binary_crossentropy` received `from_logits=True`, but the `output` argument was produced by a sigmoid or softmax activation and thus does not represent logits. Was this intended?" return dispatch_target(*args, **kwargs) 938/938 [==============================] - 31s 33ms/step - loss: 0.4817 - accuracy: 0.7691 Epoch 2/10 938/938 [==============================] - 30s 32ms/step - loss: 0.4550 - accuracy: 0.7875 Epoch 3/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4512 - accuracy: 0.7911 Epoch 4/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4487 - accuracy: 0.7940 Epoch 5/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4466 - accuracy: 0.7934 Epoch 6/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4459 - accuracy: 0.7933 Epoch 7/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4448 - accuracy: 0.7935 Epoch 8/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4439 - accuracy: 0.7950 Epoch 9/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4421 - accuracy: 0.7956 Epoch 10/10 938/938 [==============================] - 31s 32ms/step - loss: 0.4425 - accuracy: 0.7962 <keras.callbacks.History at 0x7fb364fd2a90>
由於僅使用了一小部分資料集，因此在訓練階段我們的準確性僅限於約 78%。但是，請隨時在 kafka 中儲存其他資料，以獲得更好的模型效能。此外，由於目標只是示範 tfio kafka 資料集的功能，因此使用了較小且較不複雜的神經網路。但是，可以增加模型的複雜性、修改學習策略、調整超參數等以進行探索。如需基準方法，請參閱這篇文章。
在測試資料上推論
為了在測試資料上進行推論，同時遵守「完全一次」語意以及容錯能力，可以使用 streaming.KafkaGroupIODataset。
定義 tfio 測試資料集
stream_timeout 參數會封鎖給定持續時間，以將新的資料點串流到主題中。如果資料以間歇方式串流到主題中，則無需建立新的資料集。
test_ds = tfio.experimental.streaming.KafkaGroupIODataset(
topics=["susy-test"],
group_id="testcg",
servers="127.0.0.1:9092",
stream_timeout=10000,
configuration=[
"session.timeout.ms=7000",
"max.poll.interval.ms=8000",
"auto.offset.reset=earliest"
],
)
def decode_kafka_test_item(raw_message, raw_key):
message = tf.io.decode_csv(raw_message, [[0.0] for i in range(NUM_COLUMNS)])
key = tf.strings.to_number(raw_key)
return (message, key)
test_ds = test_ds.map(decode_kafka_test_item)
test_ds = test_ds.batch(BATCH_SIZE)
WARNING:tensorflow:From /tmpfs/src/tf_docs_env/lib/python3.7/site-packages/tensorflow_io/python/experimental/kafka_group_io_dataset_ops.py:188: take_while (from tensorflow.python.data.experimental.ops.take_while_ops) is deprecated and will be removed in a future version. Instructions for updating: Use `tf.data.Dataset.take_while(...)
雖然此類別可用於訓練目的，但有些注意事項需要解決。一旦從 kafka 讀取所有訊息，並且使用 streaming.KafkaGroupIODataset 提交最新的偏移量，消費者就不會從頭開始重新讀取訊息。因此，在訓練期間，僅可能使用持續流入的資料訓練一個 epoch。這種功能在訓練階段的使用案例有限，其中一旦模型消耗了資料點，就不再需要該資料點，並且可以丟棄。
但是，當涉及到具有完全一次語意的穩健推論時，此功能會發光發熱。
評估測試資料的效能
res = model.evaluate(test_ds)
print("test loss, test acc:", res)
34/Unknown - 0s 2ms/step - loss: 0.4434 - accuracy: 0.8194 2022-01-07 20:34:29.402707: E tensorflow_io/core/kernels/kafka_kernels.cc:774] REBALANCE: Local: Assign partitions 2022-01-07 20:34:29.406789: E tensorflow_io/core/kernels/kafka_kernels.cc:776] Retrieved committed offsets with status code: 0 625/625 [==============================] - 11s 17ms/step - loss: 0.4437 - accuracy: 0.7915 test loss, test acc: [0.4436523914337158, 0.7915250062942505] 2022-01-07 20:34:40.051954: E tensorflow_io/core/kernels/kafka_kernels.cc:1001] Local: Timed out
由於推論基於「完全一次」語意，因此測試集的評估只能執行一次。為了再次在測試資料上執行推論，應使用新的消費者群組。
追蹤 testcg 消費者群組的偏移量延遲
./kafka_2.13-2.7.2/bin/kafka-consumer-groups.sh --bootstrap-server 127.0.0.1:9092 --describe --group testcg
GROUP TOPIC PARTITION CURRENT-OFFSET LOG-END-OFFSET LAG CONSUMER-ID HOST CLIENT-ID testcg susy-test 0 21626 21626 0 rdkafka-534f63d0-b91e-4976-a3ca-832b6c91210e /10.142.0.103 rdkafka testcg susy-test 1 18374 18374 0 rdkafka-534f63d0-b91e-4976-a3ca-832b6c91210e /10.142.0.103 rdkafka
一旦所有分割區的 current-offset 與 log-end-offset 相符，就表示消費者已完成從 kafka 主題擷取所有訊息。
線上學習
線上機器學習範例與訓練機器學習模型的傳統/慣用方式略有不同。在前一種情況下，一旦新的資料點可用，模型就會繼續以增量方式學習/更新其參數，並且此過程預計會無限期地持續下去。這與後一種方法不同，在後一種方法中，資料集是固定的，並且模型會對其迭代 n 次。在線上學習中，模型消耗的資料可能無法再次用於訓練。
透過使用 streaming.KafkaBatchIODataset，現在可以以這種方式訓練模型。讓我們繼續使用我們的 SUSY 資料集來示範此功能。
用於線上學習的 tfio 訓練資料集
streaming.KafkaBatchIODataset 在其 API 中類似於 streaming.KafkaGroupIODataset。此外，建議使用 stream_timeout 參數來設定資料集在逾時之前封鎖新訊息的持續時間。在以下執行個體中，資料集配置了 stream_timeout 為 10000 毫秒。這表示，在從主題消耗所有訊息後，資料集將額外等待 10 秒，然後逾時並與 kafka 叢集斷線。如果在逾時之前將新訊息串流到主題中，則資料消耗和模型訓練將針對這些新消耗的資料點恢復。若要無限期地封鎖，請將其設定為 -1。
online_train_ds = tfio.experimental.streaming.KafkaBatchIODataset(
topics=["susy-train"],
group_id="cgonline",
servers="127.0.0.1:9092",
stream_timeout=10000, # in milliseconds, to block indefinitely, set it to -1.
configuration=[
"session.timeout.ms=7000",
"max.poll.interval.ms=8000",
"auto.offset.reset=earliest"
],
)
online_train_ds 產生的每個項目本身都是一個 tf.data.Dataset。因此，所有標準轉換都可以照常應用。
def decode_kafka_online_item(raw_message, raw_key):
message = tf.io.decode_csv(raw_message, [[0.0] for i in range(NUM_COLUMNS)])
key = tf.strings.to_number(raw_key)
return (message, key)
for mini_ds in online_train_ds:
mini_ds = mini_ds.shuffle(buffer_size=32)
mini_ds = mini_ds.map(decode_kafka_online_item)
mini_ds = mini_ds.batch(32)
if len(mini_ds) > 0:
model.fit(mini_ds, epochs=3)
2022-01-07 20:34:42.024915: E tensorflow_io/core/kernels/kafka_kernels.cc:774] REBALANCE: Local: Assign partitions 2022-01-07 20:34:42.025797: E tensorflow_io/core/kernels/kafka_kernels.cc:776] Retrieved committed offsets with status code: 0 Epoch 1/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4561 - accuracy: 0.7909 Epoch 2/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4538 - accuracy: 0.7909 Epoch 3/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4499 - accuracy: 0.7947 Epoch 1/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4347 - accuracy: 0.8018 Epoch 2/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4314 - accuracy: 0.8048 Epoch 3/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4286 - accuracy: 0.8063 Epoch 1/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4480 - accuracy: 0.7910 Epoch 2/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4425 - accuracy: 0.7945 Epoch 3/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4390 - accuracy: 0.7970 Epoch 1/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4434 - accuracy: 0.7965 Epoch 2/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4380 - accuracy: 0.7974 Epoch 3/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4354 - accuracy: 0.7992 Epoch 1/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4522 - accuracy: 0.7909 Epoch 2/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4475 - accuracy: 0.7910 Epoch 3/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4435 - accuracy: 0.7947 Epoch 1/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4464 - accuracy: 0.7906 Epoch 2/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4467 - accuracy: 0.7922 Epoch 3/3 313/313 [==============================] - 1s 2ms/step - loss: 0.4424 - accuracy: 0.7933 2022-01-07 20:35:04.916208: E tensorflow_io/core/kernels/kafka_kernels.cc:1001] Local: Timed out
可以定期 (根據使用案例) 儲存以增量方式訓練的模型，並且可以用於以線上或離線模式在測試資料上進行推論。
參考文獻
Baldi, P., P. Sadowski, 和 D. Whiteson。「使用深度學習在高能物理中搜尋奇異粒子。」Nature Communications 5 (2014 年 7 月 2 日)