Check your progress

The following image shows the project Assets tab. You are now ready to visualize the training data.

Task 2a: Visualize the data

To preview this task, watch the video beginning at 01:24.

First, you visualize the data to see if there are any obvious anomalies. Follow these steps to visualize the data:

1. Open the bank-marketing-data.csv data set. The columns contain information about potential customers.
2. Click the View asset info icon to close the About this asset pane.
3. Click the Visualization tab.
   1. Select the current_outcome column as the Columns to visualize. This column indicates if the person contacted accepted the current offer; this column is the target column when you build the model.
   2. Click Add another column, and then select the previous_outcome column. This column indicates if the person contacted accepted the previous offer. Notice that the chart types with a blue dot are the suggested charts.
   3. Select the Tree chart type. This chart shows a comparison of people who accepted a previous offer and whether those same people accepted the current offer.
   4. Hover over the root node to see the total number of records in the data set.
   5. Similarly, hover over the no and yes nodes to see those totals.
   6. Hover over the unlabeled node to see the total number of records in the data set where the current_outcome field is null.

Task 2b: Prepare the data

To preview this task, watch the video beginning at 02:40.

Since current_outcome is the target column, you need to prepare the data set to remove the rows with null values. Follow these steps to prepare the data set:

1. Click Prepare data to open Data Refinery.
2. Close the About this asset pane.
3. Select the current_outcome column.
4. Click New step > Remove empty rows.
5. Click Apply.
6. Click the Profile tab to verify the action.
7. Scroll to the current_outcome column to see that all of the values are either "yes" or "no".

Task 2c: Save the refined data

To preview this task, watch the video beginning at 03:08.

To save the refined data, you specify the file name for the target data set, and then create and run a job. Follow these steps to save the refined data:

1. Specify the file name:

   1. Click the Settings icon .
   2. Click the Target data set page.
   3. Click Edit properties.
   4. Change the Data asset name* to bank-marketing-data-prepared.csv.
   5. Click Save.
   6. Click Apply.

2. Create and run the job:

   1. From the toolbar, click the Jobs icon, and select Save and create a job.
      

   2. Type the following name, Bank marketing data, and click Next.

   3. On the Configure page, select a runtime environment, and click Next.

   4. On the Schedule page, accept the default values, and click Next.

   5. On the Notify page, keep notifications turned off for this job, and click Next.

   6. Review the details, and click Create and run to run the job immediately.

3. When the job is created, click the job details link in the notification to view the job in your project. Alternatively, you can navigate to the Jobs tab in the project, and click the job name to open it.

4. When the Status for the job is Completed, use the project navigation trail to navigate back to the Assets tab in the project.

5. Click the Data > Data assets section to see the output of the Data Refinery flow, and open the bank-marketing-data-prepared.csv to verify the values.

Check your progress

The following image shows the model.

Task 4a: Promote the test data to a deployment space

To preview this task, watch the video beginning at 05:34.

The sample project includes test data. You promote that test data to a deployment space, so you can use the test data to test the deployed model. Follow these steps to promote the test data to a deployment space:

1. Return to the project's Assets tab.

2. Click the Overflow menu for the bank-marketing-test-data.csv data asset, and choose Promote to space.

3. Choose an existing deployment space. If you don't have a deployment space:

   1. Click Create a new deployment space.

   2. For the name, type: Bank marketing promotion space

   3. Select a storage service.

   4. Select a machine learning service.

   5. Click Create.

   6. Close the notification when the space is ready.

4. Select your new deployment space from the list.

5. Click Promote.

6. Click Close to return to the project's Assets tab.

Check your progress

The following image shows the completed notebook. You now have the chunked text to use to prompt the foundation model.

Task 6a: Summarization task

To preview this task, watch the video beginning at 09:32.

First, try the summarization prompt template.

1. Return to the project's Assets tab.

2. Click the Summarize Bank Promotions prompt template. This opens the prompt template in the Prompt Lab.

3. Click Edit to open the prompt template in edit mode.

   For the summarization task, you use the chunked news article text as the input example, and notes that the marketing professional usually manually writes to explain promotional offers as the output example. This is to make sure that the output is similar to what the marketing team might write themselves.

4. Notice the foundation model used for this prompt is XX.

5. Click Generate to see the summary results.

6. Experiment with different input and output text from the chunked news article in the notebook.

Task 6b: Question answering task

To preview this task, watch the video beginning at 10:15.

Next, try the question answering prompt template.

1. Click the Saved prompts to see saved prompt from your project.

2. Click the Question Answer Bank Promotions prompt template from the list of saved prompts.

3. Click Edit to open the prompt template in edit mode.

   For the question-answering task, you use questions as the input example, and answers in the level of detail required and preferred format as the output example.

4. Notice the foundation model used for this prompt is flan-t5-xxl-11b.

5. Click Generate to see the summary results.

6. Experiment with different input and output text.

Task 6c: Adjust the model parameters

To preview this task, watch the video beginning at 10:37.

In the Prompt Lab, you can adjust the decoding settings to optimize the model's output for the specific task:

• Decoding
  • Greedy: always select words with the highest probability
  • Sampling: customize the variability of word selection
• Repetition Penalty: how much repetition is allowed
• Stopping Criteria: one or more strings that will cause the text generation to stop if produced

This flexibility allows for a high degree of customization, making sure that the model operates with parameters best suited to the task's requirements and constraints.

In the Prompt Lab, you can set token limitations to make sure that the tasks remain within the operational scope of the model. This setting helps balance the response's comprehensiveness with the technical limitations of the model, resulting in an efficient and effective processing of tasks.

Task 6d: Chat with documents

To preview this task, watch the video beginning at 10:45.

You can also use Chat mode to prompt a foundation model using a sample PDF document that provides facts. Adding one or more documents creates an in-memory vector index that the foundation model will use to answer the question.

1. Click New prompt.
2. Click Chat.
3. Optional: Select a model, for example, llama-3-1-8b-instruct.
4. Click Upload documents icon , and choose Add documents.
5. Click Select from project.
   1. Select Bank CD Rates.pdf.
   2. Click Select.
6. Back on the Ground gen AI with vectorized documents page, verify the name, and click Create.
7. In the Type something field, type Which bank has the best APY for a term deposit and what is the rate?, and click the Send icon to see the answer.

Check your progress

The following image shows the Prompt Lab.