- Within the Workspace, select the Workspace item in the menu and navigate to the folder where you uploaded the Databricks Archive (which should be [your-name/AI-lab]), and select the notebook called 02 Deploy Summarizer Web Service. This will open the notebook so you can read and execute the code it contains.
- Read the instructions at the top of the notebook, and execute the cells as instructed. Remember you can use SHIFT + ENTER to execute the currently selected cell, and if you do not have a cluster attached, you will be prompted to attach to the cluster you recently deployed. Note that you will be directed to attach the azureml-sdk[databricks] library in the notebook using the same procedure you followed in Task 1.
CHAPTER 2: Create and Deploy an Unsupervised Model In this exercise, you will create and deploy a web service that uses a pre-trained model to summarize long paragraphs of text. Task 1: Install libraries The notebook you will run depends on certain Python libraries like nltk and gensim that will need to be installed in your cluster. The following steps walk you thru adding these dependencies.
Note: There are interface updates being deployed, if you do not see the Attach New button, instead go to the Azure Databricks menu option in your Workspace (the very top option on the left) and select Import Library. Then select a source of Upload Python Egg or PyPi and then provide the Package name specified in the following steps in the PyPi Name text box. Then in the Status on running clusters list, check the checkbox Attach that is listed to the left of your cluster's name to install the library on your cluster. When successful the Status should read Attached.
|
Archives
June 2020
Categories
All
|