Container Service for Kubernetes:Panduan developer Cloud Native AI Suite

Latar Belakang

Suite AI cloud-native menyediakan komponen yang dapat diterapkan secara independen (chart Helm K8s) untuk mempercepat rekayasa AI.

Prasyarat

Pastikan administrator telah menyelesaikan tugas-tugas berikut:

• Kluster Kubernetes telah dibuat.

  • Setiap node dalam kluster memiliki ruang disk minimal 300 GB.

  • Untuk hasil akselerasi data terbaik, gunakan empat instans, masing-masing dilengkapi delapan GPU V100.

  • Untuk hasil kesadaran topologi terbaik, gunakan dua instans V100.

• Suite AI cloud-native telah diinstal.

• Anda memiliki akses ke AI Dashboard.

• Anda memiliki akses ke AI Developer Console.

  Catatan

  Mulai 22 Januari 2025, AI Console Alibaba Cloud, yang mencakup AI Developer Console dan AI Dashboard, akan menjadi fitur hanya untuk daftar putih. Perubahan ini tidak memengaruhi penerapan yang sudah ada. Pengguna yang tidak masuk daftar putih dapat menginstal dan mengonfigurasi AI Console dari komunitas open-source. Untuk detailnya, lihat Open-source AI Console.

• Dataset Fashion MNIST telah diunduh dan diunggah ke OSS.

• Anda memiliki URL repositori Git, username, dan password untuk kode uji.

• Klien kubectl telah terhubung ke kluster.

• Arena CLI telah dikonfigurasi.

Lingkungan tutorial

Tujuan

Langkah 1: Buat akun dan alokasikan sumber daya

Langkah 2: Buat dataset

Dataset dikelola oleh administrator. Contoh ini menggunakan dataset fashion-mnist.

Langkah 1: Buat dataset fashion-mnist

1. Buat file fashion-mnist.yaml berdasarkan contoh YAML berikut.

   Contoh ini membuat volume persisten (PV) dan klaim volume persisten (PVC) bertipe OSS.

   apiVersion: v1
   kind: PersistentVolume
   metadata:
     name: fashion-demo-pv
   spec:
     accessModes:
     - ReadWriteMany
     capacity:
       storage: 10Gi
     csi:
       driver: ossplugin.csi.alibabacloud.com
       volumeAttributes:
         bucket: fashion-mnist
         otherOpts: ""
         url: oss-cn-beijing.aliyuncs.com
         akId: "AKID"
         akSecret: "AKSECRET"
       volumeHandle: fashion-demo-pv
     persistentVolumeReclaimPolicy: Retain
     storageClassName: oss
     volumeMode: Filesystem
   ---
   apiVersion: v1
   kind: PersistentVolumeClaim
   metadata:
     name: fashion-demo-pvc
     namespace: demo-ns
   spec:
     accessModes:
     - ReadWriteMany
     resources:
       requests:
         storage: 10Gi
     selector:
       matchLabels:
         alicloud-pvname: fashion-demo-pv
     storageClassName: oss
     volumeMode: Filesystem
     volumeName: fashion-demo-pv

2. Buat dataset fashion-mnist.

   kubectl create -f fashion-mnist.yaml

3. Lihat status PV dan PVC.

   • Periksa status PV.

     kubectl get pv fashion-mnist-jackwg

     Output yang diharapkan:

     NAME                   CAPACITY   ACCESS MODES   RECLAIM POLICY   STATUS   CLAIM                          STORAGECLASS   REASON   AGE
     fashion-mnist-jackwg   10Gi       RWX            Retain           Bound    ns1/fashion-mnist-jackwg-pvc   oss                     8h

   • Periksa status PVC.

     kubectl get pvc fashion-mnist-jackwg-pvc -n ns1

     Output yang diharapkan:

     NAME                       STATUS   VOLUME                 CAPACITY   ACCESS MODES   STORAGECLASS   AGE
     fashion-mnist-jackwg-pvc   Bound    fashion-mnist-jackwg   10Gi       RWX            oss            8h

   Baik PV maupun PVC menunjukkan status Bound.

Langkah 2: Buat dataset terakselerasi

Administrator menggunakan AI Dashboard untuk mempercepat dataset.

1. Akses AI Dashboard sebagai administrator.
2. Di panel navigasi kiri AI Dashboard, pilih Dataset > Dataset List.

3. Pada halaman Daftar Dataset, klik Accelerate di kolom Aksi dataset target.

   Setelah Anda mempercepat dataset, informasinya diperbarui di halaman Daftar Dataset. Daftar tersebut mencakup kolom seperti Nama Tugas, Namespace, Sumber Data, Dipercepat, dan Status. Untuk dataset yang dipercepat, kolom Dipercepat menampilkan Yes, dan Status-nya berubah dari NotReady menjadi Ready setelah akselerasi selesai.

Langkah 3: Kembangkan model

(Opsional) Langkah 1: Buat notebook dari gambar kustom

AI Developer Console menyediakan Jupyter Notebook dengan gambar bawaan untuk berbagai versi TensorFlow dan PyTorch. Jika gambar tersebut tidak memenuhi kebutuhan Anda, Anda dapat membuat gambar kustom.

1. Buat file bernama Dockerfile berdasarkan templat berikut.

   Lihat Buat dan gunakan notebook untuk persyaratan gambar kustom.

   cat<<EOF >dockerfile
   FROM tensorflow/tensorflow:1.15.5-gpu
   USER root
   RUN pip install jupyter && \
       pip install ipywidgets && \
       jupyter nbextension enable --py widgetsnbextension && \
       pip install jupyterlab && jupyter serverextension enable --py jupyterlab
   EXPOSE 8888
   #USER jovyan
   CMD ["sh", "-c", "jupyter-lab --notebook-dir=/home/jovyan --ip=0.0.0.0 --no-browser --allow-root --port=8888 --NotebookApp.token='' --NotebookApp.password='' --NotebookApp.allow_origin='*' --NotebookApp.base_url=${NB_PREFIX} --ServerApp.authenticate_prometheus=False"]
   EOF

2. Bangun gambar dari Dockerfile.

   docker build -f dockerfile .

   Output yang diharapkan:

   Sending build context to Docker daemon  9.216kB
   Step 1/5 : FROM tensorflow/tensorflow:1.15.5-gpu
    ---> 73be11373498
   Step 2/5 : USER root
    ---> Using cache
    ---> 7ee21dc7e42e
   Step 3/5 : RUN pip install jupyter &&     pip install ipywidgets &&     jupyter nbextension enable --py widgetsnbextension &&     pip install jupyterlab && jupyter serverextension enable --py jupyterlab
    ---> Using cache
    ---> 23bc51c5e16d
   Step 4/5 : EXPOSE 8888
    ---> Using cache
    ---> 76a55822ddae
   Step 5/5 : CMD ["sh", "-c", "jupyter-lab --notebook-dir=/home/jovyan --ip=0.0.0.0 --no-browser --allow-root --port=8888 --NotebookApp.token='' --NotebookApp.password='' --NotebookApp.allow_origin='*' --NotebookApp.base_url=${NB_PREFIX} --ServerApp.authenticate_prometheus=False"]
    ---> Using cache
    ---> 3692f04626d5
   Successfully built 3692f04626d5

3. Dorong gambar ke registri kontainer Anda.

   docker tag ${IMAGE_ID} registry-vpc.cn-beijing.aliyuncs.com/${DOCKER_REPO}/jupyter:fashion-mnist-20210802a
   docker push registry-vpc.cn-beijing.aliyuncs.com/${DOCKER_REPO}/jupyter:fashion-mnist-20210802a

4. Buat rahasia tarik gambar untuk menarik gambar dari registri kontainer pribadi Anda.

   Lihat Buat Secret untuk menarik gambar dari registri pribadi.

   kubectl create secret docker-registry regcred \
     --docker-server=<your-registry-server> \
     --docker-username=<your-username> \
     --docker-password=<your-password> \
     --docker-email=<your-email-address>

5. Buat Jupyter Notebook di AI Developer Console.

   Lihat Buat dan gunakan notebook.

   Konfigurasikan parameter untuk Jupyter Notebook sebagai berikut: Di bagian Informasi Dasar Notebook di sebelah kiri, atur Nama Notebook dan Gambar Notebook (Anda dapat memilih gambar kustom), serta tentukan Namespace dan Rahasia Tarik Gambar. Untuk Konfigurasi Data, pilih dataset untuk menampilkan klaim volume persisten dan direktori penyimpanan lokal yang sesuai. Jika Anda mengaktifkan Workspace PVC, Anda harus menentukan PVC Target dan Jalur Mount. Di bagian Konfigurasi Sumber Daya Notebook di sebelah kanan, atur CPU (Core), Memori (GB), dan GPU (Unit). Setelah selesai mengonfigurasi, klik Create Notebook.

Langkah 2: Kembangkan dan uji di notebook

1. Akses AI Developer Console
   
2. Di panel navigasi kiri AI Developer Console, klik Notebook.

3. Di halaman Notebook, klik nama Jupyter Notebook target dengan Status Running.

4. Buka terminal baru dan jalankan perintah berikut untuk memverifikasi bahwa data berhasil dimount:

   pwd
   /root/data
   ls -alh

   Output yang diharapkan:

   total 30M
   drwx------ 1 root root    0 Jan  1  1970 .
   drwx------ 1 root root 4.0K Aug  2 04:15 ..
   drwxr-xr-x 1 root root    0 Aug  1 14:16 saved_model
   -rw-r----- 1 root root 4.3M Aug  1 01:53 t10k-images-idx3-ubyte.gz
   -rw-r----- 1 root root 5.1K Aug  1 01:53 t10k-labels-idx1-ubyte.gz
   -rw-r----- 1 root root  26M Aug  1 01:54 train-images-idx3-ubyte.gz
   -rw-r----- 1 root root  29K Aug  1 01:53 train-labels-idx1-ubyte.gz

5. Di lingkungan Jupyter Notebook Anda, buat file notebook baru untuk mengembangkan model fashion-mnist dan inisialisasi dengan kode berikut:

   #!/usr/bin/python
   # -*- coding: UTF-8 -*-
   
   import os
   import gzip
   import numpy as np
   import tensorflow as tf
   from tensorflow import keras
   print('TensorFlow version: {}'.format(tf.__version__))
   dataset_path = "/root/data/"
   model_path = "./model/"
   model_version =  "v1"
   
   def load_data():
       files = [
           'train-labels-idx1-ubyte.gz',
           'train-images-idx3-ubyte.gz',
           't10k-labels-idx1-ubyte.gz',
           't10k-images-idx3-ubyte.gz'
       ]
       paths = []
       for fname in files:
           paths.append(os.path.join(dataset_path, fname))
       with gzip.open(paths[0], 'rb') as labelpath:
           y_train = np.frombuffer(labelpath.read(), np.uint8, offset=8)
       with gzip.open(paths[1], 'rb') as imgpath:
           x_train = np.frombuffer(imgpath.read(), np.uint8, offset=16).reshape(len(y_train), 28, 28)
       with gzip.open(paths[2], 'rb') as labelpath:
           y_test = np.frombuffer(labelpath.read(), np.uint8, offset=8)
       with gzip.open(paths[3], 'rb') as imgpath:
           x_test = np.frombuffer(imgpath.read(), np.uint8, offset=16).reshape(len(y_test), 28, 28)
       return (x_train, y_train),(x_test, y_test)
   
   def train():
       (train_images, train_labels), (test_images, test_labels) = load_data()
   
       # scale the values to 0.0 to 1.0
       train_images = train_images / 255.0
       test_images = test_images / 255.0
   
       # reshape for feeding into the model
       train_images = train_images.reshape(train_images.shape[0], 28, 28, 1)
       test_images = test_images.reshape(test_images.shape[0], 28, 28, 1)
   
       class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',
                   'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']
   
       print('\ntrain_images.shape: {}, of {}'.format(train_images.shape, train_images.dtype))
       print('test_images.shape: {}, of {}'.format(test_images.shape, test_images.dtype))
   
       model = keras.Sequential([
       keras.layers.Conv2D(input_shape=(28,28,1), filters=8, kernel_size=3,
                           strides=2, activation='relu', name='Conv1'),
       keras.layers.Flatten(),
       keras.layers.Dense(10, activation=tf.nn.softmax, name='Softmax')
       ])
       model.summary()
       testing = False
       epochs = 5
       model.compile(optimizer='adam',
                   loss='sparse_categorical_crossentropy',
                   metrics=['accuracy'])
       logdir = "/training_logs"
       tensorboard_callback = keras.callbacks.TensorBoard(log_dir=logdir)
       model.fit(train_images,
           train_labels,
           epochs=epochs,
           callbacks=[tensorboard_callback],
       )
       test_loss, test_acc = model.evaluate(test_images, test_labels)
       print('\nTest accuracy: {}'.format(test_acc))
       export_path = os.path.join(model_path, model_version)
       print('export_path = {}\n'.format(export_path))
       tf.keras.models.save_model(
           model,
           export_path,
           overwrite=True,
           include_optimizer=True,
           save_format=None,
           signatures=None,
           options=None
       )
       print('\nSaved model success')
   if __name__ == '__main__':
       train()

   Penting

   Dalam kode tersebut, dataset_path dan model_path harus diatur ke jalur mount sumber data di Notebook. Hal ini memungkinkan Notebook mengakses dataset yang dimount ke sistem file lokal.

6. Di Notebook target, klik ikon untuk menjalankan kode.

   Output yang diharapkan:

   TensorFlow version: 1.15.5
   
   train_images.shape: (60000, 28, 28, 1), of float64
   test_images.shape: (10000, 28, 28, 1), of float64
   Model: "sequential_2"
   _________________________________________________________________
   Layer (type)                 Output Shape              Param #
   =================================================================
   Conv1 (Conv2D)               (None, 13, 13, 8)         80
   _________________________________________________________________
   flatten_2 (Flatten)          (None, 1352)              0
   _________________________________________________________________
   Softmax (Dense)              (None, 10)                13530
   =================================================================
   Total params: 13,610
   Trainable params: 13,610
   Non-trainable params: 0
   _________________________________________________________________
   Train on 60000 samples
   Epoch 1/5
   60000/60000 [==============================] - 3s 57us/sample - loss: 0.5452 - acc: 0.8102
   Epoch 2/5
   60000/60000 [==============================] - 3s 52us/sample - loss: 0.4103 - acc: 0.8555
   Epoch 3/5
   60000/60000 [==============================] - 3s 55us/sample - loss: 0.3750 - acc: 0.8681
   Epoch 4/5
   60000/60000 [==============================] - 3s 55us/sample - loss: 0.3524 - acc: 0.8757
   Epoch 5/5
   60000/60000 [==============================] - 3s 53us/sample - loss: 0.3368 - acc: 0.8798
   10000/10000 [==============================] - 0s 37us/sample - loss: 0.3770 - acc: 0.8673
   
   Test accuracy: 0.8672999739646912
   export_path = ./model/v1
   Saved model success

Langkah 3: Dorong kode ke repositori Git

Dorong kode ke repositori Git Anda langsung dari Notebook.

1. Instal Git.

   apt-get update
   apt-get install git

2. Inisialisasi konfigurasi Git dan simpan kredensial Anda.

   git config --global credential.helper store
   git pull ${YOUR_GIT_REPO}

3. Dorong kode ke repositori Git.

   git push origin fashion-test

   Output yang diharapkan:

   Total 0 (delta 0), reused 0 (delta 0)
   To codeup.aliyun.com:60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git
    * [new branch]      fashion-test -> fashion-test

Langkah 4: Kirimkan job pelatihan menggunakan Arena SDK

1. Instal dependensi untuk Arena SDK.

   !pip install coloredlogs

2. Buat file Python baru dan inisialisasi dengan kode berikut:

   import os
   import sys
   import time
   from arenasdk.client.client import ArenaClient
   from arenasdk.enums.types import *
   from arenasdk.exceptions.arena_exception import *
   from arenasdk.training.tensorflow_job_builder import *
   from arenasdk.logger.logger import LoggerBuilder
   
   def main():
       print("start to test arena-python-sdk")
       client = ArenaClient("","demo-ns","info","arena-system") # Job dikirimkan ke namespace demo-ns.
       print("create ArenaClient succeed.")
       print("start to create tfjob")
       job_name = "arena-sdk-distributed-test"
       job_type = TrainingJobType.TFTrainingJob
       try:
           # bangun job pelatihan
           job =  TensorflowJobBuilder().with_name(job_name)\
               .with_workers(1)\
               .with_gpus(1)\
               .with_worker_image("tensorflow/tensorflow:1.5.0-devel-gpu")\
               .with_ps_image("tensorflow/tensorflow:1.5.0-devel")\
               .with_ps_count(1)\
               .with_datas({"fashion-demo-pvc":"/data"})\
               .enable_tensorboard()\
               .with_sync_mode("git")\
               .with_sync_source("https://codeup.aliyun.com/60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git")\  # URL repositori Git.
               .with_envs({\
                   "GIT_SYNC_USERNAME":"USERNAME", \   # Username untuk repositori Git.
                   "GIT_SYNC_PASSWORD":"PASSWORD",\    # Password untuk repositori Git.
                   "TEST_TMPDIR":"/",\
               })\
               .with_command("python code/tensorflow-fashion-mnist-sample/tf-distributed-mnist.py").build()
           # Jika job pelatihan sudah ada, hapus
           if client.training().get(job_name, job_type):
               print("The job {} already exists. Deleting it.".format(job_name))
               client.training().delete(job_name, job_type)
               time.sleep(3)
   
           output = client.training().submit(job)
           print(output)
   
           count = 0
           # Tunggu hingga job pelatihan mulai berjalan.
           while True:
               if count > 160:
                   raise Exception("Timeout waiting for the job to start running")
               jobInfo = client.training().get(job_name,job_type)
               if jobInfo.get_status() == TrainingJobStatus.TrainingJobPending:
                   print("job status is PENDING,waiting...")
                   count = count + 1
                   time.sleep(5)
                   continue
               print("current status is {} of job {}".format(jobInfo.get_status().value,job_name))
               break
           # dapatkan log job pelatihan
           logger = LoggerBuilder().with_accepter(sys.stdout).with_follow().with_since("5m")
           #jobInfo.get_instances()[0].get_logs(logger)
           # tampilkan informasi job pelatihan
           print(str(jobInfo))
           # hapus job pelatihan
           #client.training().delete(job_name, job_type)
       except ArenaException as e:
           print(e)
   
   main()

   • namespace: Dalam contoh ini, job pelatihan dikirimkan ke namespace demo-ns.

   • with_sync_source: URL repositori Git.

   • with_envs: Username dan password untuk repositori Git.

3. Di Notebook target, klik ikon untuk menjalankan kode.

   Output yang diharapkan:

   2021-11-02/08:57:28 DEBUG util.py[line:19] - execute command: [arena get --namespace=demo-ns --arena-namespace=arena-system --loglevel=info arena-sdk-distributed-test --type=tfjob -o json]
   2021-11-02/08:57:28 DEBUG util.py[line:19] - execute command: [arena submit --namespace=demo-ns --arena-namespace=arena-system --loglevel=info tfjob --name=arena-sdk-distributed-test --workers=1 --gpus=1 --worker-image=tensorflow/tensorflow:1.5.0-devel-gpu --ps-image=tensorflow/tensorflow:1.5.0-devel --ps=1 --data=fashion-demo-pvc:/data --tensorboard --sync-mode=git --sync-source=https://codeup.aliyun.com/60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git --env=GIT_SYNC_USERNAME=kubeai --env=GIT_SYNC_PASSWORD=kubeai@ACK123 --env=TEST_TMPDIR=/ python code/tensorflow-fashion-mnist-sample/tf-distributed-mnist.py]
   start to test arena-python-sdk
   create ArenaClient succeed.
   start to create tfjob
   2021-11-02/08:57:29 DEBUG util.py[line:19] - execute command: [arena get --namespace=demo-ns --arena-namespace=arena-system --loglevel=info arena-sdk-distributed-test --type=tfjob -o json]
   service/arena-sdk-distributed-test-tensorboard created
   deployment.apps/arena-sdk-distributed-test-tensorboard created
   tfjob.kubeflow.org/arena-sdk-distributed-test created
   
   job status is PENDING,waiting...
   2021-11-02/09:00:34 DEBUG util.py[line:19] - execute command: [arena get --namespace=demo-ns --arena-namespace=arena-system --loglevel=info arena-sdk-distributed-test --type=tfjob -o json]
   current status is RUNNING of job arena-sdk-distributed-test
   {
       "allocated_gpus": 1,
       "chief_name": "arena-sdk-distributed-test-worker-0",
       "duration": "185s",
       "instances": [
           {
               "age": "13s",
               "gpu_metrics": [],
               "is_chief": false,
               "name": "arena-sdk-distributed-test-ps-0",
               "node_ip": "192.168.5.8",
               "node_name": "cn-beijing.192.168.5.8",
               "owner": "arena-sdk-distributed-test",
               "owner_type": "tfjob",
               "request_gpus": 0,
               "status": "Running"
           },
           {
               "age": "13s",
               "gpu_metrics": [],
               "is_chief": true,
               "name": "arena-sdk-distributed-test-worker-0",
               "node_ip": "192.168.5.8",
               "node_name": "cn-beijing.192.168.5.8",
               "owner": "arena-sdk-distributed-test",
               "owner_type": "tfjob",
               "request_gpus": 1,
               "status": "Running"
           }
       ],
       "name": "arena-sdk-distributed-test",
       "namespace": "demo-ns",
       "priority": "N/A",
       "request_gpus": 1,
       "tensorboard": "http://192.168.5.6:31068",
       "type": "tfjob"
   }

Langkah 4: Latih model

Contoh-contoh berikut mencakup pelatihan standalone, pelatihan terdistribusi, pelatihan dipercepat Fluid, dan penjadwalan sadar topologi dengan penjadwal job AI.

Contoh 1: Kirimkan job pelatihan TensorFlow standalone

Kirimkan job pelatihan dengan Arena CLI atau AI Developer Console.

Metode 1: Gunakan Arena CLI

arena \
  submit \
  tfjob \
  -n ns1 \
  --name=fashion-mnist-arena \
  --data=fashion-mnist-jackwg-pvc:/root/data/ \
  --env=DATASET_PATH=/root/data/ \
  --env=MODEL_PATH=/root/saved_model \
  --env=MODEL_VERSION=1 \
  --env=GIT_SYNC_USERNAME=<GIT_USERNAME> \
  --env=GIT_SYNC_PASSWORD=<GIT_PASSWORD> \
  --sync-mode=git \
  --sync-source=https://codeup.aliyun.com/60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git \
  --image="tensorflow/tensorflow:2.2.2-gpu" \
  "python /root/code/tensorflow-fashion-mnist-sample/train.py --log_dir=/training_logs"

Metode 2: Gunakan AI Developer Console

1. Konfigurasikan sumber data.

   Tabel berikut menjelaskan beberapa parameter utama.

   Parameter	Contoh	Wajib
   Name	fashion-demo	Ya
   Namespace	demo-ns	Ya
   PersistentVolumeClaim	fashion-demo-pvc	Ya
   Direktori penyimpanan lokal	/root/data	Tidak

2. Konfigurasikan repositori kode sumber.

   Parameter	Contoh	Wajib
   Name	fashion-git	Ya
   Git URL	https://codeup.aliyun.com/60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git	Ya
   Default branch	master	Tidak
   Direktori penyimpanan lokal	/root/	Tidak
   Username Git pribadi	Username untuk repositori Git pribadi Anda.	Tidak
   Password atau token Git pribadi	Password atau personal access token untuk repositori Git pribadi Anda.	Tidak

3. Kirimkan job pelatihan standalone.

   Setelah Anda mengonfigurasi parameter pelatihan, klik Submit. Anda kemudian dapat melihat job pelatihan di halaman Daftar Job. Tabel berikut menjelaskan parameter pengiriman job.

   Parameter	Deskripsi
   Task Name	Dalam contoh ini, nama job adalah fashion-tf-ui.
   Job type	Untuk contoh ini, pilih Tensorflow Standalone.
   Namespace	Dalam contoh ini, gunakan demo-ns. Ini harus namespace yang sama dengan dataset.
   Konfigurasi sumber data	Untuk contoh ini, pilih fashion-demo, yang dikonfigurasi di Langkah 1.
   Konfigurasi kode	Untuk contoh ini, pilih fashion-git, yang dikonfigurasi di Langkah 2.
   Code branch	Dalam contoh ini, gunakan master.
   Command	Dalam contoh ini, gunakan "export DATASET_PATH=/root/data/ &&export MODEL_PATH=/root/saved_model &&export MODEL_VERSION=1 &&python /root/code/tensorflow-fashion-mnist-sample/train.py".
   Repositori Git pribadi	Jika Anda menggunakan repositori pribadi, Anda harus memberikan username dan password.
   Jumlah instans	Nilai default adalah 1.
   Image	Dalam contoh ini, gunakan tensorflow/tensorflow:2.2.2-gpu.
   Image pull secret	Jika Anda menggunakan gambar dari registri pribadi, Anda harus membuat rahasia tarik gambar terlebih dahulu.
   CPU (inti)	Nilai default adalah 4.
   Memori (GB)	Nilai default adalah 8.

   Lihat Parameter Arena CLI untuk TFJob.

4. Setelah mengirimkan job, lihat log-nya.

   1. Di panel navigasi kiri AI Developer Console, klik Tasks.

   2. Di halaman Tasks, klik nama job target.

   3. Di halaman detail job, klik tab Instance. Lalu, di kolom Actions untuk instans target, klik Logs.

      Cuplikan berikut menunjukkan output log untuk contoh ini:

      train_images.shape: (60000, 28, 28, 1), of float64
      test_images.shape: (10000, 28, 28, 1), of float64
      Model: "sequential"
      _________________________________________________________________
      Layer (type)                 Output Shape              Param #
      =================================================================
      Conv1 (Conv2D)               (None, 13, 13, 8)         80
      _________________________________________________________________
      flatten (Flatten)            (None, 1352)              0
      _________________________________________________________________
      Softmax (Dense)              (None, 10)                13530
      =================================================================
      Total params: 13,610
      Trainable params: 13,610
      Non-trainable params: 0
      _________________________________________________________________
      Epoch 1/5
      2021-08-01 14:21:17.532237: E tensorflow/core/profiler/internal/gpu/cupti_tracer.cc:1430] function cupti_interface_->EnableCallback( 0 , subscriber_, CUPTI_CB_DOMAIN_DRIVER_API, cbid)failed with error CUPTI_ERROR_INVALID_PARAMETER
      2021-08-01 14:21:17.532390: I tensorflow/core/profiler/internal/gpu/device_tracer.cc:216]  GpuTracer has collected 0 callback api events and 0 activity events.
      2021-08-01 14:21:17.533535: I tensorflow/core/profiler/rpc/client/save_profile.cc:168] Creating directory: /training_logs/train/plugins/profile/2021_08_01_14_21_17
      2021-08-01 14:21:17.533928: I tensorflow/core/profiler/rpc/client/save_profile.cc:174] Dumped gzipped tool data for trace.json.gz to /training_logs/train/plugins/profile/2021_08_01_14_21_17/fashion-mnist-arena-chief-0.trace.json.gz
      2021-08-01 14:21:17.534251: I tensorflow/core/profiler/utils/event_span.cc:288] Generation of step-events took 0 ms
      
      2021-08-01 14:21:17.534961: I tensorflow/python/profiler/internal/profiler_wrapper.cc:87] Creating directory: /training_logs/train/plugins/profile/2021_08_01_14_21_17Dumped tool data for overview_page.pb to /training_logs/train/plugins/profile/2021_08_01_14_21_17/fashion-mnist-arena-chief-0.overview_page.pb
      Dumped tool data for input_pipeline.pb to /training_logs/train/plugins/profile/2021_08_01_14_21_17/fashion-mnist-arena-chief-0.input_pipeline.pb
      Dumped tool data for tensorflow_stats.pb to /training_logs/train/plugins/profile/2021_08_01_14_21_17/fashion-mnist-arena-chief-0.tensorflow_stats.pb
      Dumped tool data for kernel_stats.pb to /training_logs/train/plugins/profile/2021_08_01_14_21_17/fashion-mnist-arena-chief-0.kernel_stats.pb
      
      1875/1875 [==============================] - 3s 2ms/step - loss: 0.5399 - accuracy: 0.8116
      Epoch 2/5
      1875/1875 [==============================] - 3s 2ms/step - loss: 0.4076 - accuracy: 0.8573
      Epoch 3/5
      1875/1875 [==============================] - 3s 2ms/step - loss: 0.3727 - accuracy: 0.8694
      Epoch 4/5
      1875/1875 [==============================] - 3s 2ms/step - loss: 0.3512 - accuracy: 0.8769
      Epoch 5/5
      1875/1875 [==============================] - 3s 2ms/step - loss: 0.3351 - accuracy: 0.8816
      313/313 [==============================] - 0s 1ms/step - loss: 0.3595 - accuracy: 0.8733
      2021-08-01 14:21:34.820089: W tensorflow/python/util/util.cc:329] Sets are not currently considered sequences, but this may change in the future, so consider avoiding using them.
      WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/resource_variable_ops.py:1817: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.
      Instructions for updating:
      If using Keras pass *_constraint arguments to layers.
      
      Test accuracy: 0.8733000159263611
      export_path = /root/saved_model/1
      
      Saved model success

5. Lihat metrik pelatihan di TensorBoard.

   Teruskan port layanan TensorBoard ke mesin lokal Anda menggunakan kubectl port-forward.

   1. Dapatkan detail layanan TensorBoard.

      kubectl get svc -n demo-ns

      Output yang diharapkan:

      NAME                        TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)               AGE
      tf-dist-arena-tensorboard   NodePort    172.16.XX.XX     <none>        6006:32226/TCP        80m

   2. Teruskan port TensorBoard.

      kubectl port-forward svc/tf-dist-arena-tensorboard -n demo-ns 6006:6006

      Output yang diharapkan:

      Forwarding from 127.0.0.1:6006 -> 6006
      Forwarding from [::1]:6006 -> 6006
      Handling connection for 6006
      Handling connection for 6006

   3. Di browser web, buka http://localhost:6006/ untuk melihat dasbor TensorBoard.

      Tab GRAPHS menampilkan graf komputasi model, menunjukkan koneksi dan aliran data antar lapisan seperti Conv1, flatten, dan Softmax. Panel samping menyediakan properti detail untuk node yang dipilih.

Contoh 2: Kirimkan job pelatihan TensorFlow terdistribusi

Metode 1: Gunakan Arena CLI

1. Kirimkan job pelatihan.

   arena submit tf \
       -n demo-ns \
       --name=tf-dist-arena \
       --working-dir=/root/ \
       --data fashion-mnist-pvc:/data \
       --env=TEST_TMPDIR=/ \
       --env=GIT_SYNC_USERNAME=kubeai \
       --env=GIT_SYNC_PASSWORD=kubeai@ACK123 \
       --env=GIT_SYNC_BRANCH=master \
       --gpus=1 \
       --workers=2 \
       --worker-image=tensorflow/tensorflow:1.5.0-devel-gpu \
       --sync-mode=git \
       --sync-source=https://codeup.aliyun.com/60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git \
       --ps=1 \
       --ps-image=tensorflow/tensorflow:1.5.0-devel \
       --tensorboard \
       "python code/tensorflow-fashion-mnist-sample/tf-distributed-mnist.py --log_dir=/training_logs"

2. Dapatkan detail layanan TensorBoard.

   kubectl get svc -n demo-ns

   Output yang diharapkan:

   NAME                        TYPE        CLUSTER-IP       EXTERNAL-IP   PORT(S)                 AGE
   tf-dist-arena-tensorboard   NodePort    172.16.204.248   <none>        6006:32226/TCP          80m

3. Teruskan port TensorBoard.

   kubectl port-forward svc/tf-dist-arena-tensorboard -n demo-ns 6006:6006

   Output yang diharapkan:

   Forwarding from 127.0.0.1:6006 -> 6006
   Forwarding from [::1]:6006 -> 6006
   Handling connection for 6006
   Handling connection for 6006

4. Di browser web, buka http://localhost:6006/ untuk melihat dasbor TensorBoard.

   Tab SCALARS menampilkan plot untuk metrik seperti accuracy dan cross_entropy selama pelatihan. Di panel samping, Anda dapat memfilter berdasarkan run train dan test, serta menyesuaikan opsi tampilan Smoothing dan sumbu.

Metode 2: Gunakan AI Developer Console

1. Konfigurasikan sumber data.

   Gunakan kembali data dari Langkah 1.

2. Konfigurasikan repositori kode sumber.

   Gunakan kembali kode dari Langkah 2.

3. Kirimkan job pelatihan TensorFlow terdistribusi.

   Setelah Anda mengonfigurasi parameter pelatihan, klik Submit. Anda kemudian dapat melihat job di halaman Daftar Job. Tabel berikut menjelaskan parameter job utama.

   Parameter	Deskripsi
   Task Name	Dalam contoh ini, gunakan fashion-ps-ui.
   Job type	Pilih TF Distributed.
   Namespace	Dalam contoh ini, gunakan demo-ns. Ini harus namespace yang sama dengan dataset.
   Konfigurasi sumber data	Dalam contoh ini, pilih fashion-demo, yang dikonfigurasi di Langkah 1 contoh sebelumnya.
   Konfigurasi kode	Dalam contoh ini, pilih fashion-git, yang dikonfigurasi di Langkah 2 contoh sebelumnya.
   Command	Dalam contoh ini, gunakan "export TEST_TMPDIR=/root/ && python code/tensorflow-fashion-mnist-sample/tf-distributed-mnist.py --log_dir=/training_logs".
   Image	Di bawah Konfigurasi Sumber Daya Tugas, pada tab Worker, atur Image ke tensorflow/tensorflow:1.5.0-devel-gpu. Di bawah Konfigurasi Sumber Daya Tugas, pada tab PS, atur Image ke tensorflow/tensorflow:1.5.0-devel.

   Lihat Parameter Arena CLI untuk TFJob.

4. Untuk melihat TensorBoard, ikuti langkah 2 hingga 4 di Metode 1: Gunakan Arena CLI.

Contoh 3: Kirimkan job pelatihan dipercepat Fluid

1. Percepat dataset yang ada.

   Lewati langkah ini jika Anda sudah mempercepat fashion-demo-pvc di Langkah 2: Buat dataset. Lihat Kelola dataset.

2. Kirimkan job yang menggunakan dataset yang dipercepat.

   Kirimkan job pelatihan ke namespace demo-ns. Perbedaan utama untuk dataset yang dipercepat:

   • --data: Nama PersistentVolumeClaim (PVC) yang dipercepat. Dalam contoh ini, yaitu fashion-demo-pvc-acc.

   • --env=DATASET_PATH: Jalur ini dibentuk dengan menambahkan nama PVC (fashion-demo-pvc-acc) ke jalur mount (/root/data/) yang ditentukan dalam parameter --data.

   arena \
     submit \
     tfjob \
     -n demo-ns \
     --name=fashion-mnist-fluid \
     --data=fashion-demo-pvc-acc:/root/data/ \
     --env=DATASET_PATH=/root/data/fashion-demo-pvc-acc \
     --env=MODEL_PATH=/root/saved_model \
     --env=MODEL_VERSION=1 \
     --env=GIT_SYNC_USERNAME=${GIT_USERNAME} \
     --env=GIT_SYNC_PASSWORD=${GIT_PASSWORD} \
     --sync-mode=git \
     --sync-source=https://codeup.aliyun.com/60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git \
     --image="tensorflow/tensorflow:2.2.2-gpu" \
     "python /root/code/tensorflow-fashion-mnist-sample/train.py --log_dir=/training_logs"

3. Bandingkan waktu eksekusi kedua job pelatihan.

   arena list -n demo-ns

   Output yang diharapkan:

   NAME                 STATUS     TRAINER  DURATION  GPU(Requested)  GPU(Allocated)  NODE
   fashion-mnist-fluid  SUCCEEDED  TFJOB    33s       0               N/A             192.168.5.7
   fashion-mnist-arena  SUCCEEDED  TFJOB    3m        0               N/A             192.168.5.8

   Dengan kode dan sumber daya yang sama, job yang dipercepat Fluid selesai dalam 33 detik dibandingkan 3 menit untuk job standar.

Contoh 4: Percepat job pelatihan terdistribusi dengan menggunakan penjadwal job AI

Penjadwal job AI adalah plugin ACK untuk workload AI dan data besar, mendukung Gang Scheduling, Capacity Scheduling, dan penjadwalan sadar topologi. Contoh ini menunjukkan penjadwalan sadar topologi GPU.

Penjadwal menggunakan informasi topologi tingkat node (tautan GPU seperti NVLink dan PCIe Switch, atau arsitektur CPU NUMA) untuk mengoptimalkan penjadwalan job AI. Lihat Penjadwalan sadar topologi GPU dan Penjadwalan sadar topologi CPU.

Aktifkan penjadwalan sadar topologi GPU dan bandingkan kinerja job.

1. Buat job pelatihan tanpa penjadwalan sadar topologi.

   arena submit mpi \
     --name=tensorflow-4-vgg16 \
     --gpus=1 \
     --workers=4 \
     --image=registry.cn-hangzhou.aliyuncs.com/kubernetes-image-hub/tensorflow-benchmark:tf2.3.0-py3.7-cuda10.1 \
     "mpirun --allow-run-as-root -np "4" -bind-to none -map-by slot -x NCCL_DEBUG=INFO -x NCCL_SOCKET_IFNAME=eth0 -x LD_LIBRARY_PATH -x PATH --mca pml ob1 --mca btl_tcp_if_include eth0 --mca oob_tcp_if_include eth0 --mca orte_keep_fqdn_hostnames t --mca btl ^openib python /tensorflow/benchmarks/scripts/tf_cnn_benchmarks/tf_cnn_benchmarks.py --model=vgg16 --batch_size=64 --variable_update=horovod"

2. Buat job pelatihan dengan penjadwalan sadar topologi diaktifkan.

   Beri label node. Ganti cn-beijing.192.168.XX.XX dengan nama node aktual Anda.

   kubectl label node cn-beijing.192.168.XX.XX ack.node.gpu.schedule=topology --overwrite

   Buat job pelatihan. Bendera --gputopology=true mengaktifkan kesadaran topologi di Arena.

   arena submit mpi \
     --name=tensorflow-topo-4-vgg16 \
     --gpus=1 \
     --workers=4 \
     --gputopology=true \
     --image=registry.cn-hangzhou.aliyuncs.com/kubernetes-image-hub/tensorflow-benchmark:tf2.3.0-py3.7-cuda10.1 \
     "mpirun --allow-run-as-root -np "4" -bind-to none -map-by slot -x NCCL_DEBUG=INFO -x NCCL_SOCKET_IFNAME=eth0 -x LD_LIBRARY_PATH -x PATH --mca pml ob1 --mca btl_tcp_if_include eth0 --mca oob_tcp_if_include eth0 --mca orte_keep_fqdn_hostnames t --mca btl ^openib python /tensorflow/benchmarks/scripts/tf_cnn_benchmarks/tf_cnn_benchmarks.py --model=vgg16 --batch_size=64 --variable_update=horovod

3. Bandingkan kinerja kedua job.

   1. Bandingkan waktu eksekusi.

      arena list -n demo-ns

      Output yang diharapkan:

      NAME                             STATUS     TRAINER  DURATION  GPU(Requested)  GPU(Allocated)  NODE
      tensorflow-topo-4-vgg16          SUCCEEDED  MPIJOB   44s       4               N/A             192.168.4.XX1
      tensorflow-4-vgg16-image-warned  SUCCEEDED  MPIJOB   2m        4               N/A             192.168.4.XX0

   2. Lihat throughput job sadar topologi.

      arena logs tensorflow-topo-4-vgg16 -n demo-ns

      Output yang diharapkan:

      100 images/sec: 251.7 +/- 0.1 (jitter = 1.2)  7.262
      ----------------------------------------------------------------
      total images/sec: 1006.44

   3. Lihat throughput job standar.

      arena logs tensorflow-4-vgg16-image-warned -n demo-ns

      Output yang diharapkan:

      100 images/sec: 56.4 +/- 0.2 (jitter = 1.5)  7.261
      ----------------------------------------------------------------
      total images/sec: 225.50

kubectl label node cn-beijing.192.168.XX.XX0 ack.node.gpu.schedule=default --overwrite

Langkah 5: Kelola model

1. Akses AI Developer Console
   
2. Di panel navigasi kiri AI Developer Console, klik Model Manage.

3. Di halaman Model Management, klik Create Model.

4. Di kotak dialog Create, konfigurasikan Model Name, Model Version, dan Job Name.

   Dalam contoh ini, nama model adalah fashion-mnist-demo, versi model adalah v1, dan job pelatihan adalah tf-single.

5. Klik OK. Model baru muncul dalam daftar.

   Untuk mengevaluasi model, klik New Model Evaluation di baris model tersebut.

Langkah 6: Evaluasi model

1. Kirimkan job pelatihan dengan checkpointing diaktifkan.

   Kirimkan job pelatihan dengan Arena. Job tersebut menghasilkan checkpoint ke volume fashion-demo-pvc.

   arena \
     submit \
     tfjob \
     -n demo-ns \ # Atur namespace sesuai kebutuhan.
     --name=fashion-mnist-arena-ckpt \
     --data=fashion-demo-pvc:/root/data/ \
     --env=DATASET_PATH=/root/data/ \
     --env=MODEL_PATH=/root/data/saved_model \
     --env=MODEL_VERSION=1 \
     --env=GIT_SYNC_USERNAME=${GIT_USERNAME} \ # Masukkan username Git Anda.
     --env=GIT_SYNC_PASSWORD=${GIT_PASSWORD} \ # Masukkan password Git Anda.
     --env=OUTPUT_CHECKPOINT=1 \
     --sync-mode=git \
     --sync-source=https://codeup.aliyun.com/60b4cf5c66bba1c04b442e49/tensorflow-fashion-mnist-sample.git \
     --image="tensorflow/tensorflow:2.2.2-gpu" \
     "python /root/code/tensorflow-fashion-mnist-sample/train.py --log_dir=/training_logs"

2. Kirimkan job evaluasi model.

   1. Bangun gambar evaluasi.

      Di direktori kubeai-sdk, jalankan perintah berikut untuk membangun dan mendorong gambar.

      docker build . -t ${DOCKER_REGISTRY}:fashion-mnist
      docker push ${DOCKER_REGISTRY}:fashion-mnist

   2. Dapatkan detail layanan MySQL.

      kubectl get svc -n kube-ai ack-mysql

      Output yang diharapkan:

      NAME        TYPE        CLUSTER-IP      EXTERNAL-IP   PORT(S)    AGE
      ack-mysql   ClusterIP   172.16.XX.XX    <none>        3306/TCP   28h

   3. Kirimkan job evaluasi model dengan Arena.

      arena evaluate model \
       --namespace=demo-ns \
       --loglevel=debug \
       --name=evaluate-job \
       --image=registry.cn-beijing.aliyuncs.com/kube-ai/kubeai-sdk-demo:fashion-minist \
       --env=ENABLE_MYSQL=True \
       --env=MYSQL_HOST=172.16.77.227 \
       --env=MYSQL_PORT=3306 \
       --env=MYSQL_USERNAME=kubeai \
       --env=MYSQL_PASSWORD=kubeai@ACK \
       --data=fashion-demo-pvc:/data \
       --model-name=1 \
       --model-path=/data/saved_model/ \
       --dataset-path=/data/ \
       --metrics-path=/data/output \
       "python /kubeai/evaluate.py"

      Catatan

      Anda dapat memperoleh alamat IP dan port layanan MySQL dari langkah sebelumnya.

3. Bandingkan hasil evaluasi.

   1. Di panel navigasi kiri AI Developer Console, klik Model Management.

      Halaman tersebut menampilkan daftar job evaluasi model. Daftar tersebut mencakup kolom seperti Name, Model name, Model version, Namespace, Status, Creation time, dan End time. Untuk job yang selesai, kolom Status menampilkan Complete. Anda dapat membandingkan hasil beberapa job dengan mengklik tombol Compare Metrics di bagian atas halaman.

   2. Di Tasks, klik nama job untuk melihat metriknya.

      Hasil evaluasi menampilkan metrik berikut: accuracy, precision, recall, F1_score, dan AUC. Kurva ROC juga memvisualisasikan kinerja klasifikasi model.

      Grafik batang membandingkan kinerja job yang dipilih berdasarkan metrik seperti AUC dan accuracy.

Langkah 7: Terapkan model

Terapkan model yang telah Anda latih sebagai layanan inferensi TensorFlow Serving. Arena juga mendukung framework lain seperti Triton dan Seldon (lihat Dokumentasi serve Arena).

Contoh ini menggunakan model dari Langkah 4, yang disimpan di PVC fashion-demo-pvc dari Langkah 2. Jika model Anda menggunakan jenis penyimpanan berbeda, buat PVC yang sesuai terlebih dahulu.

1. Terapkan model TensorFlow ke TensorFlow Serving dengan Arena.

   arena serve tensorflow \
     --loglevel=debug \
     --namespace=demo-ns \
     --name=fashion-mnist \
     --model-name=1  \
     --gpus=1  \
     --image=tensorflow/serving:1.15.0-gpu \
     --data=fashion-demo-pvc:/data \
     --model-path=/data/saved_model/ \
     --version-policy=latest

2. Dapatkan nama layanan inferensi yang diterapkan.

   arena serve list -n demo-ns

   Output yang diharapkan:

   NAME           TYPE        VERSION       DESIRED  AVAILABLE  ADDRESS         PORTS                   GPU
   fashion-mnist  Tensorflow  202111031203  1        1          172.16.XX.XX    GRPC:8500,RESTFUL:8501  1

   Anda dapat menggunakan nilai ADDRESS dan PORTS dalam output untuk memanggil layanan dari dalam kluster.

3. Di Jupyter, buat file Notebook baru untuk mengirim permintaan ke layanan HTTP TensorFlow Serving.

   Contoh ini menggunakan Jupyter Notebook yang dibuat di Langkah 3: Kembangkan model untuk mengirim permintaan.

   • Dalam kode inisialisasi berikut, ganti nilai server_ip dengan ADDRESS (172.16.XX.XX) yang dikembalikan pada langkah sebelumnya.

   • Atur server_http_port ke port RESTFUL (8501) yang dikembalikan pada langkah sebelumnya.

   Kode inisialisasi untuk file Notebook adalah sebagai berikut:

   import os
   import gzip
   import numpy as np
   # import matplotlib.pyplot as plt
   import random
   import requests
   import json
   
   server_ip = "172.16.XX.XX"
   server_http_port = 8501
   
   dataset_dir = "/root/data/"
   
   def load_data():
           files = [
               'train-labels-idx1-ubyte.gz',
               'train-images-idx3-ubyte.gz',
               't10k-labels-idx1-ubyte.gz',
               't10k-images-idx3-ubyte.gz'
           ]
   
           paths = []
           for fname in files:
               paths.append(os.path.join(dataset_dir, fname))
   
           with gzip.open(paths[0], 'rb') as labelpath:
               y_train = np.frombuffer(labelpath.read(), np.uint8, offset=8)
           with gzip.open(paths[1], 'rb') as imgpath:
               x_train = np.frombuffer(imgpath.read(), np.uint8, offset=16).reshape(len(y_train), 28, 28)
           with gzip.open(paths[2], 'rb') as labelpath:
               y_test = np.frombuffer(labelpath.read(), np.uint8, offset=8)
           with gzip.open(paths[3], 'rb') as imgpath:
               x_test = np.frombuffer(imgpath.read(), np.uint8, offset=16).reshape(len(y_test), 28, 28)
   
           return (x_train, y_train),(x_test, y_test)
   
   def show(idx, title):
     plt.figure()
     plt.imshow(test_images[idx].reshape(28,28))
     plt.axis('off')
     plt.title('\n\n{}'.format(title), fontdict={'size': 16})
   
   class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',
                  'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']
   
   (train_images, train_labels), (test_images, test_labels) = load_data()
   train_images = train_images / 255.0
   test_images = test_images / 255.0
   
   # reshape for feeding into the model
   train_images = train_images.reshape(train_images.shape[0], 28, 28, 1)
   test_images = test_images.reshape(test_images.shape[0], 28, 28, 1)
   
   print('\ntrain_images.shape: {}, of {}'.format(train_images.shape, train_images.dtype))
   print('test_images.shape: {}, of {}'.format(test_images.shape, test_images.dtype))
   
   rando = random.randint(0,len(test_images)-1)
   #show(rando, 'An Example Image: {}'.format(class_names[test_labels[rando]]))
   
   # !pip install -q requests
   
   # import requests
   # headers = {"content-type": "application/json"}
   # json_response = requests.post('http://localhost:8501/v1/models/fashion_model:predict', data=data, headers=headers)
   # predictions = json.loads(json_response.text)['predictions']
   
   # show(0, 'The model thought this was a {} (class {}), and it was actually a {} (class {})'.format(
   #   class_names[np.argmax(predictions[0])], np.argmax(predictions[0]), class_names[test_labels[0]], test_labels[0]))
   
   def request_model(data):
       headers = {"content-type": "application/json"}
       json_response = requests.post('http://{}:{}/v1/models/1:predict'.format(server_ip, server_http_port), data=data, headers=headers)
       print('=======response:', json_response, json_response.text)
       predictions = json.loads(json_response.text)['predictions']
   
       print('The model thought this was a {} (class {}), and it was actually a {} (class {})'.format(class_names[np.argmax(predictions[0])], np.argmax(predictions[0]), class_names[test_labels[0]], test_labels[0]))
       #show(0, 'The model thought this was a {} (class {}), and it was actually a {} (class {})'.format(
       #  class_names[np.argmax(predictions[0])], np.argmax(predictions[0]), class_names[test_labels[0]], test_labels[0]))
   
   # def request_model_version(data):
   #     headers = {"content-type": "application/json"}
   #     json_response = requests.post('http://{}:{}/v1/models/1/version/1:predict'.format(server_ip, server_http_port), data=data, headers=headers)
   #     print('=======response:', json_response, json_response.text)
   
   #     predictions = json.loads(json_response.text)
   #     for i in range(0,3):
   #       show(i, 'The model thought this was a {} (class {}), and it was actually a {} (class {})'.format(
   #         class_names[np.argmax(predictions[i])], np.argmax(predictions[i]), class_names[test_labels[i]], test_labels[i]))
   
   data = json.dumps({"signature_name": "serving_default", "instances": test_images[0:3].tolist()})
   print('Data: {} ... {}'.format(data[:50], data[len(data)-52:]))
   #request_model_version(data)
   request_model(data)

   Klik ikon di Jupyter Notebook untuk melihat output berikut:

   train_images.shape: (60000, 28, 28, 1), of float64
   test_images.shape: (10000, 28, 28, 1), of float64
   Data: {"signature_name": "serving_default", "instances": ...  [0.0], [0.0], [0.0], [0.0], [0.0], [0.0], [0.0]]]]}
   =======response: <Response [200]> {
       "predictions": [[7.42696e-07, 6.91237556e-09, 2.66364452e-07, 2.27735413e-07, 4.0373439e-07, 0.00490919966, 7.27086217e-06, 0.0316713452, 0.0010733594, 0.962337255], [0.00685342, 1.8516447e-08, 0.9266119, 2.42278338e-06, 0.0603800081, 4.01338771e-12, 0.00613868702, 4.26091073e-15, 1.35764185e-05, 3.38685469e-10], [1.09047969e-05, 0.999816835, 7.98738e-09, 0.000122893631, 4.85748023e-05, 1.50353979e-10, 3.57102294e-07, 1.89657579e-09, 4.4604468e-07, 9.23274524e-09]
       ]
   }
   The model thought this was a Ankle boot (class 9), and it was actually a Ankle boot (class 9)

FAQ

• Bagaimana cara menginstal perangkat lunak di konsol Jupyter Notebook?

  Untuk menginstal perangkat lunak, jalankan perintah berikut di konsol Jupyter Notebook.

  apt-get install <software-name>

• Bagaimana cara memperbaiki karakter acak di konsol Jupyter Notebook?

  Edit file /etc/locale agar berisi hal berikut, lalu buka kembali terminal.

  LC_CTYPE="da_DK.UTF-8"
  LC_NUMERIC="da_DK.UTF-8"
  LC_TIME="da_DK.UTF-8"
  LC_COLLATE="da_DK.UTF-8"
  LC_MONETARY="da_DK.UTF-8"
  LC_MESSAGES="da_DK.UTF-8"
  LC_PAPER="da_DK.UTF-8"
  LC_NAME="da_DK.UTF-8"
  LC_ADDRESS="da_DK.UTF-8"
  LC_TELEPHONE="da_DK.UTF-8"
  LC_MEASUREMENT="da_DK.UTF-8"
  LC_IDENTIFICATION="da_DK.UTF-8"
  LC_ALL=