🎨 Data Designer Tutorial: Structured Outputs, Jinja Expressions, and Conditional Generation¶

📚 What you'll learn¶

In this notebook, we will continue our exploration of Data Designer, demonstrating more advanced data generation using structured outputs, Jinja expressions, and conditional generation with skip.when.

If this is your first time using Data Designer, we recommend starting with the first notebook in this tutorial series.

📦 Import Data Designer¶

• data_designer.config provides access to the configuration API.

• DataDesigner is the main interface for data generation.

In [1]:

import data_designer.config as dd
from data_designer.interface import DataDesigner

import data_designer.config as dd from data_designer.interface import DataDesigner

⚙️ Initialize the Data Designer interface¶

• DataDesigner is the main object that is used to interface with the library.

• When initialized without arguments, the default model providers are used.

In [2]:

data_designer = DataDesigner()

data_designer = DataDesigner()

🎛️ Define model configurations¶

• Each ModelConfig defines a model that can be used during the generation process.

• The "model alias" is used to reference the model in the Data Designer config (as we will see below).

• The "model provider" is the external service that hosts the model (see the model config docs for more details).

• By default, we use build.nvidia.com as the model provider.

In [3]:

# This name is set in the model provider configuration.
MODEL_PROVIDER = "nvidia"

# The model ID is from build.nvidia.com.
MODEL_ID = "nvidia/nemotron-3-nano-30b-a3b"

# We choose this alias to be descriptive for our use case.
MODEL_ALIAS = "nemotron-nano-v3"

model_configs = [
    dd.ModelConfig(
        alias=MODEL_ALIAS,
        model=MODEL_ID,
        provider=MODEL_PROVIDER,
        inference_parameters=dd.ChatCompletionInferenceParams(
            temperature=1.0,
            top_p=1.0,
            max_tokens=2048,
            extra_body={"chat_template_kwargs": {"enable_thinking": False}},
        ),
    )
]

# This name is set in the model provider configuration. MODEL_PROVIDER = "nvidia" # The model ID is from build.nvidia.com. MODEL_ID = "nvidia/nemotron-3-nano-30b-a3b" # We choose this alias to be descriptive for our use case. MODEL_ALIAS = "nemotron-nano-v3" model_configs = [ dd.ModelConfig( alias=MODEL_ALIAS, model=MODEL_ID, provider=MODEL_PROVIDER, inference_parameters=dd.ChatCompletionInferenceParams( temperature=1.0, top_p=1.0, max_tokens=2048, extra_body={"chat_template_kwargs": {"enable_thinking": False}}, ), ) ]

🏗️ Initialize the Data Designer Config Builder¶

• The Data Designer config defines the dataset schema and generation process.

• The config builder provides an intuitive interface for building this configuration.

• The list of model configs is provided to the builder at initialization.

In [4]:

config_builder = dd.DataDesignerConfigBuilder(model_configs=model_configs)

config_builder = dd.DataDesignerConfigBuilder(model_configs=model_configs)

🧑‍🎨 Designing our data¶

• We will again create a product review dataset, but this time we will use structured outputs and Jinja expressions.

• Structured outputs let you specify the exact schema of the data you want to generate.

• Data Designer supports schemas specified using either json schema or Pydantic data models (recommended).

We'll define our structured outputs using Pydantic data models

💡 Why Pydantic?

• Pydantic models provide better IDE support and type validation.

• They are more Pythonic than raw JSON schemas.

• They integrate seamlessly with Data Designer's structured output system.

In [5]:

from decimal import Decimal
from typing import Literal

from pydantic import BaseModel, Field


# We define a Product schema so that the name, description, and price are generated
# in one go, with the types and constraints specified.
class Product(BaseModel):
    name: str = Field(description="The name of the product")
    description: str = Field(description="A description of the product")
    price: Decimal = Field(description="The price of the product", ge=10, le=1000, decimal_places=2)


class ProductReview(BaseModel):
    rating: int = Field(description="The rating of the product", ge=1, le=5)
    customer_mood: Literal["irritated", "mad", "happy", "neutral", "excited"] = Field(
        description="The mood of the customer"
    )
    review: str = Field(description="A review of the product")

from decimal import Decimal from typing import Literal from pydantic import BaseModel, Field # We define a Product schema so that the name, description, and price are generated # in one go, with the types and constraints specified. class Product(BaseModel): name: str = Field(description="The name of the product") description: str = Field(description="A description of the product") price: Decimal = Field(description="The price of the product", ge=10, le=1000, decimal_places=2) class ProductReview(BaseModel): rating: int = Field(description="The rating of the product", ge=1, le=5) customer_mood: Literal["irritated", "mad", "happy", "neutral", "excited"] = Field( description="The mood of the customer" ) review: str = Field(description="A review of the product")

Next, let's design our product review dataset using a few more tricks compared to the previous notebook.

In [6]:

# Since we often only want a few attributes from Person objects, we can
# set drop=True in the column config to drop the column from the final dataset.
config_builder.add_column(
    dd.SamplerColumnConfig(
        name="customer",
        sampler_type=dd.SamplerType.PERSON_FROM_FAKER,
        params=dd.PersonFromFakerSamplerParams(),
        drop=True,
    )
)

config_builder.add_column(
    dd.SamplerColumnConfig(
        name="product_category",
        sampler_type=dd.SamplerType.CATEGORY,
        params=dd.CategorySamplerParams(
            values=[
                "Electronics",
                "Clothing",
                "Home & Kitchen",
                "Books",
                "Home Office",
            ],
        ),
    )
)

config_builder.add_column(
    dd.SamplerColumnConfig(
        name="product_subcategory",
        sampler_type=dd.SamplerType.SUBCATEGORY,
        params=dd.SubcategorySamplerParams(
            category="product_category",
            values={
                "Electronics": [
                    "Smartphones",
                    "Laptops",
                    "Headphones",
                    "Cameras",
                    "Accessories",
                ],
                "Clothing": [
                    "Men's Clothing",
                    "Women's Clothing",
                    "Winter Coats",
                    "Activewear",
                    "Accessories",
                ],
                "Home & Kitchen": [
                    "Appliances",
                    "Cookware",
                    "Furniture",
                    "Decor",
                    "Organization",
                ],
                "Books": [
                    "Fiction",
                    "Non-Fiction",
                    "Self-Help",
                    "Textbooks",
                    "Classics",
                ],
                "Home Office": [
                    "Desks",
                    "Chairs",
                    "Storage",
                    "Office Supplies",
                    "Lighting",
                ],
            },
        ),
    )
)

config_builder.add_column(
    dd.SamplerColumnConfig(
        name="target_age_range",
        sampler_type=dd.SamplerType.CATEGORY,
        params=dd.CategorySamplerParams(values=["18-25", "25-35", "35-50", "50-65", "65+"]),
    )
)

# Sampler columns support conditional params, which are used if the condition is met.
# In this example, we set the review style to rambling if the target age range is 18-25.
# Note conditional parameters are only supported for Sampler column types.
config_builder.add_column(
    dd.SamplerColumnConfig(
        name="review_style",
        sampler_type=dd.SamplerType.CATEGORY,
        params=dd.CategorySamplerParams(
            values=["rambling", "brief", "detailed", "structured with bullet points"],
            weights=[1, 2, 2, 1],
        ),
        conditional_params={
            "target_age_range == '18-25'": dd.CategorySamplerParams(values=["rambling"]),
        },
    )
)

# Optionally validate that the columns are configured correctly.
data_designer.validate(config_builder)

# Since we often only want a few attributes from Person objects, we can # set drop=True in the column config to drop the column from the final dataset. config_builder.add_column( dd.SamplerColumnConfig( name="customer", sampler_type=dd.SamplerType.PERSON_FROM_FAKER, params=dd.PersonFromFakerSamplerParams(), drop=True, ) ) config_builder.add_column( dd.SamplerColumnConfig( name="product_category", sampler_type=dd.SamplerType.CATEGORY, params=dd.CategorySamplerParams( values=[ "Electronics", "Clothing", "Home & Kitchen", "Books", "Home Office", ], ), ) ) config_builder.add_column( dd.SamplerColumnConfig( name="product_subcategory", sampler_type=dd.SamplerType.SUBCATEGORY, params=dd.SubcategorySamplerParams( category="product_category", values={ "Electronics": [ "Smartphones", "Laptops", "Headphones", "Cameras", "Accessories", ], "Clothing": [ "Men's Clothing", "Women's Clothing", "Winter Coats", "Activewear", "Accessories", ], "Home & Kitchen": [ "Appliances", "Cookware", "Furniture", "Decor", "Organization", ], "Books": [ "Fiction", "Non-Fiction", "Self-Help", "Textbooks", "Classics", ], "Home Office": [ "Desks", "Chairs", "Storage", "Office Supplies", "Lighting", ], }, ), ) ) config_builder.add_column( dd.SamplerColumnConfig( name="target_age_range", sampler_type=dd.SamplerType.CATEGORY, params=dd.CategorySamplerParams(values=["18-25", "25-35", "35-50", "50-65", "65+"]), ) ) # Sampler columns support conditional params, which are used if the condition is met. # In this example, we set the review style to rambling if the target age range is 18-25. # Note conditional parameters are only supported for Sampler column types. config_builder.add_column( dd.SamplerColumnConfig( name="review_style", sampler_type=dd.SamplerType.CATEGORY, params=dd.CategorySamplerParams( values=["rambling", "brief", "detailed", "structured with bullet points"], weights=[1, 2, 2, 1], ), conditional_params={ "target_age_range == '18-25'": dd.CategorySamplerParams(values=["rambling"]), }, ) ) # Optionally validate that the columns are configured correctly. data_designer.validate(config_builder)

[03:29:18] [INFO] ✅ Validation passed

Next, we will use more advanced Jinja expressions to create new columns.

Jinja expressions let you:

• Access nested attributes: {{ customer.first_name }}

• Combine values: {{ customer.first_name }} {{ customer.last_name }}

• Use conditional logic: {% if condition %}...{% endif %}

In [7]:

# We can create new columns using Jinja expressions that reference
# existing columns, including attributes of nested objects.
config_builder.add_column(
    dd.ExpressionColumnConfig(name="customer_name", expr="{{ customer.first_name }} {{ customer.last_name }}")
)

config_builder.add_column(dd.ExpressionColumnConfig(name="customer_age", expr="{{ customer.age }}"))

config_builder.add_column(
    dd.LLMStructuredColumnConfig(
        name="product",
        prompt=(
            "Create a product in the '{{ product_category }}' category, focusing on products  "
            "related to '{{ product_subcategory }}'. The target age range of the ideal customer is "
            "{{ target_age_range }} years old. The product should be priced between $10 and $1000."
        ),
        output_format=Product,
        model_alias=MODEL_ALIAS,
    )
)

# We can even use if/else logic in our Jinja expressions to create more complex prompt patterns.
config_builder.add_column(
    dd.LLMStructuredColumnConfig(
        name="customer_review",
        prompt=(
            "Your task is to write a review for the following product:\n\n"
            "Product Name: {{ product.name }}\n"
            "Product Description: {{ product.description }}\n"
            "Price: {{ product.price }}\n\n"
            "Imagine your name is {{ customer_name }} and you are from {{ customer.city }}, {{ customer.state }}. "
            "Write the review in a style that is '{{ review_style }}'."
            "{% if target_age_range == '18-25' %}"
            "Make sure the review is more informal and conversational.\n"
            "{% else %}"
            "Make sure the review is more formal and structured.\n"
            "{% endif %}"
            "The review field should contain only the review, no other text."
        ),
        output_format=ProductReview,
        model_alias=MODEL_ALIAS,
    )
)

data_designer.validate(config_builder)

# We can create new columns using Jinja expressions that reference # existing columns, including attributes of nested objects. config_builder.add_column( dd.ExpressionColumnConfig(name="customer_name", expr="{{ customer.first_name }} {{ customer.last_name }}") ) config_builder.add_column(dd.ExpressionColumnConfig(name="customer_age", expr="{{ customer.age }}")) config_builder.add_column( dd.LLMStructuredColumnConfig( name="product", prompt=( "Create a product in the '{{ product_category }}' category, focusing on products " "related to '{{ product_subcategory }}'. The target age range of the ideal customer is " "{{ target_age_range }} years old. The product should be priced between $10 and $1000." ), output_format=Product, model_alias=MODEL_ALIAS, ) ) # We can even use if/else logic in our Jinja expressions to create more complex prompt patterns. config_builder.add_column( dd.LLMStructuredColumnConfig( name="customer_review", prompt=( "Your task is to write a review for the following product:\n\n" "Product Name: {{ product.name }}\n" "Product Description: {{ product.description }}\n" "Price: {{ product.price }}\n\n" "Imagine your name is {{ customer_name }} and you are from {{ customer.city }}, {{ customer.state }}. " "Write the review in a style that is '{{ review_style }}'." "{% if target_age_range == '18-25' %}" "Make sure the review is more informal and conversational.\n" "{% else %}" "Make sure the review is more formal and structured.\n" "{% endif %}" "The review field should contain only the review, no other text." ), output_format=ProductReview, model_alias=MODEL_ALIAS, ) ) data_designer.validate(config_builder)

[03:29:18] [INFO] ✅ Validation passed

🚦 Conditional generation with skip.when¶

So far, every column is generated for every row. But sometimes an expensive LLM column only makes sense for a subset of rows — for example, a detailed complaint analysis is only useful when the review is negative.

Data Designer lets you skip column generation on a per-row basis using SkipConfig. Skipped rows receive None by default, but you can provide a sentinel value with skip=dd.SkipConfig(when="...", value="N/A") to write a specific value instead.

There are three patterns to know:

Pattern	How	Effect
Expression gate	skip=dd.SkipConfig(when="...")	Skip this column when the Jinja2 expression is truthy
Skip propagation (default)	Downstream column depends on a skipped column	Automatically skipped too (propagate_skip=True by default)
Propagation opt-out	propagate_skip=False on the downstream column	Always generates, even if an upstream was skipped

Pattern 1 — Expression gate. Only generate a detailed complaint analysis when the customer gave a low rating (1 or 2 stars). Rows where the rating is 3 or higher will get None for this column.

In [8]:

config_builder.add_column(
    dd.LLMTextColumnConfig(
        name="complaint_analysis",
        model_alias=MODEL_ALIAS,
        prompt=(
            "A customer reviewed '{{ product.name }}' ({{ product_category }} / {{ product_subcategory }}).\n\n"
            "Review: {{ customer_review.review }}\n"
            "Rating: {{ customer_review.rating }}/5\n"
            "Mood: {{ customer_review.customer_mood }}\n\n"
            "Write a short root-cause analysis of why this customer is unhappy "
            "and suggest one concrete improvement the product team could make."
        ),
        skip=dd.SkipConfig(when="{{ customer_review.rating > 2 }}"),
    )
)

config_builder.add_column( dd.LLMTextColumnConfig( name="complaint_analysis", model_alias=MODEL_ALIAS, prompt=( "A customer reviewed '{{ product.name }}' ({{ product_category }} / {{ product_subcategory }}).\n\n" "Review: {{ customer_review.review }}\n" "Rating: {{ customer_review.rating }}/5\n" "Mood: {{ customer_review.customer_mood }}\n\n" "Write a short root-cause analysis of why this customer is unhappy " "and suggest one concrete improvement the product team could make." ), skip=dd.SkipConfig(when="{{ customer_review.rating > 2 }}"), ) )

Out[8]:

DataDesignerConfigBuilder(
    sampler_columns: [
        "customer",
        "product_category",
        "product_subcategory",
        "target_age_range",
        "review_style"
    ]
    llm_text_columns: ['complaint_analysis']
    llm_structured_columns: ['product', 'customer_review']
    expression_columns: ['customer_name', 'customer_age']
)

Pattern 2 — Skip propagation. action_items depends on complaint_analysis. When complaint_analysis is skipped, action_items auto-skips too because propagate_skip defaults to True.

In [9]:

config_builder.add_column(
    dd.LLMTextColumnConfig(
        name="action_items",
        model_alias=MODEL_ALIAS,
        prompt=(
            "Based on this complaint analysis:\n"
            "{{ complaint_analysis }}\n\n"
            "List 2-3 concrete action items for the product team."
        ),
    )
)

config_builder.add_column( dd.LLMTextColumnConfig( name="action_items", model_alias=MODEL_ALIAS, prompt=( "Based on this complaint analysis:\n" "{{ complaint_analysis }}\n\n" "List 2-3 concrete action items for the product team." ), ) )

Out[9]:

DataDesignerConfigBuilder(
    sampler_columns: [
        "customer",
        "product_category",
        "product_subcategory",
        "target_age_range",
        "review_style"
    ]
    llm_text_columns: ['complaint_analysis', 'action_items']
    llm_structured_columns: ['product', 'customer_review']
    expression_columns: ['customer_name', 'customer_age']
)

Pattern 3 — Propagation opt-out. review_summary also depends on complaint_analysis, but sets propagate_skip=False so it always generates. The prompt uses a Jinja conditional to handle the case where complaint_analysis is None.

In [10]:

config_builder.add_column(
    dd.LLMTextColumnConfig(
        name="review_summary",
        model_alias=MODEL_ALIAS,
        propagate_skip=False,
        prompt=(
            "Summarize this product review in one sentence:\n"
            "Product: {{ product.name }}\n"
            "Rating: {{ customer_review.rating }}/5\n"
            "Review: {{ customer_review.review }}\n"
            "{% if complaint_analysis %}"
            "Complaint analysis: {{ complaint_analysis }}\n"
            "{% endif %}"
        ),
    )
)

data_designer.validate(config_builder)

config_builder.add_column( dd.LLMTextColumnConfig( name="review_summary", model_alias=MODEL_ALIAS, propagate_skip=False, prompt=( "Summarize this product review in one sentence:\n" "Product: {{ product.name }}\n" "Rating: {{ customer_review.rating }}/5\n" "Review: {{ customer_review.review }}\n" "{% if complaint_analysis %}" "Complaint analysis: {{ complaint_analysis }}\n" "{% endif %}" ), ) ) data_designer.validate(config_builder)

[03:29:18] [INFO] ✅ Validation passed

🔁 Iteration is key – preview the dataset!¶

1. Use the preview method to generate a sample of records quickly.

2. Inspect the results for quality and format issues.

3. Adjust column configurations, prompts, or parameters as needed.

4. Re-run the preview until satisfied.

In [11]:

preview = data_designer.preview(config_builder, num_records=2)

preview = data_designer.preview(config_builder, num_records=2)

[03:29:18] [INFO] 🔁 Preview generation in progress

[03:29:18] [INFO]   |-- 🔒 Jinja rendering engine: secure

[03:29:18] [INFO] ✅ Validation passed

[03:29:18] [INFO] ⛓️ Sorting column configs into a Directed Acyclic Graph

[03:29:18] [INFO] 🩺 Running health checks for models...

[03:29:18] [INFO]   |-- 👀 Checking 'nvidia/nemotron-3-nano-30b-a3b' in provider named 'nvidia' for model alias 'nemotron-nano-v3'...

[03:29:19] [INFO]   |-- ✅ Passed!

[03:29:19] [INFO] 🎲 Preparing samplers to generate 2 records across 5 columns

[03:29:19] [INFO] 🧩 Generating column `customer_name` from expression

[03:29:19] [INFO] 🧩 Generating column `customer_age` from expression

[03:29:19] [INFO] 🗂️ llm-structured model config for column 'product'

[03:29:19] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:19] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:19] [INFO]   |-- model provider: 'nvidia'

[03:29:19] [INFO]   |-- inference parameters:

[03:29:19] [INFO]   |  |-- generation_type=chat-completion

[03:29:19] [INFO]   |  |-- max_parallel_requests=4

[03:29:19] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:19] [INFO]   |  |-- temperature=1.00

[03:29:19] [INFO]   |  |-- top_p=1.00

[03:29:19] [INFO]   |  |-- max_tokens=2048

[03:29:19] [INFO] ⚡️ Processing llm-structured column 'product' with 4 concurrent workers

[03:29:19] [INFO] ⏱️ llm-structured column 'product' will report progress after each record

[03:29:20] [INFO]   |-- 🐥 llm-structured column 'product' progress: 1/2 (50%) complete, 1 ok, 0 failed, 1.25 rec/s, eta 0.8s

[03:29:21] [INFO]   |-- 🐔 llm-structured column 'product' progress: 2/2 (100%) complete, 2 ok, 0 failed, 1.86 rec/s, eta 0.0s

[03:29:21] [INFO] 🗂️ llm-structured model config for column 'customer_review'

[03:29:21] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:21] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:21] [INFO]   |-- model provider: 'nvidia'

[03:29:21] [INFO]   |-- inference parameters:

[03:29:21] [INFO]   |  |-- generation_type=chat-completion

[03:29:21] [INFO]   |  |-- max_parallel_requests=4

[03:29:21] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:21] [INFO]   |  |-- temperature=1.00

[03:29:21] [INFO]   |  |-- top_p=1.00

[03:29:21] [INFO]   |  |-- max_tokens=2048

[03:29:21] [INFO] ⚡️ Processing llm-structured column 'customer_review' with 4 concurrent workers

[03:29:21] [INFO] ⏱️ llm-structured column 'customer_review' will report progress after each record

[03:29:22] [INFO]   |-- 😸 llm-structured column 'customer_review' progress: 1/2 (50%) complete, 1 ok, 0 failed, 0.57 rec/s, eta 1.8s

[03:29:24] [INFO]   |-- 🦁 llm-structured column 'customer_review' progress: 2/2 (100%) complete, 2 ok, 0 failed, 0.66 rec/s, eta 0.0s

[03:29:24] [INFO] 📝 llm-text model config for column 'complaint_analysis'

[03:29:24] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:24] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:24] [INFO]   |-- model provider: 'nvidia'

[03:29:24] [INFO]   |-- inference parameters:

[03:29:24] [INFO]   |  |-- generation_type=chat-completion

[03:29:24] [INFO]   |  |-- max_parallel_requests=4

[03:29:24] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:24] [INFO]   |  |-- temperature=1.00

[03:29:24] [INFO]   |  |-- top_p=1.00

[03:29:24] [INFO]   |  |-- max_tokens=2048

[03:29:24] [INFO] ⚡️ Processing llm-text column 'complaint_analysis' with 4 concurrent workers

[03:29:24] [INFO] ⏱️ llm-text column 'complaint_analysis' will report progress after each record

[03:29:24] [INFO]   |-- 🚗 llm-text column 'complaint_analysis' progress: 1/2 (50%) complete, 0 ok, 0 failed, 1 skipped, 522.80 rec/s, eta 0.0s

[03:29:24] [INFO]   |-- 🚀 llm-text column 'complaint_analysis' progress: 2/2 (100%) complete, 0 ok, 0 failed, 2 skipped, 779.08 rec/s, eta 0.0s

[03:29:24] [INFO] 📝 llm-text model config for column 'action_items'

[03:29:24] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:24] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:24] [INFO]   |-- model provider: 'nvidia'

[03:29:24] [INFO]   |-- inference parameters:

[03:29:24] [INFO]   |  |-- generation_type=chat-completion

[03:29:24] [INFO]   |  |-- max_parallel_requests=4

[03:29:24] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:24] [INFO]   |  |-- temperature=1.00

[03:29:24] [INFO]   |  |-- top_p=1.00

[03:29:24] [INFO]   |  |-- max_tokens=2048

[03:29:24] [INFO] ⚡️ Processing llm-text column 'action_items' with 4 concurrent workers

[03:29:24] [INFO] ⏱️ llm-text column 'action_items' will report progress after each record

[03:29:24] [INFO]   |-- 😐 llm-text column 'action_items' progress: 1/2 (50%) complete, 0 ok, 0 failed, 1 skipped, 1190.83 rec/s, eta 0.0s

[03:29:24] [INFO]   |-- 🤩 llm-text column 'action_items' progress: 2/2 (100%) complete, 0 ok, 0 failed, 2 skipped, 1642.23 rec/s, eta 0.0s

[03:29:24] [INFO] 📝 llm-text model config for column 'review_summary'

[03:29:24] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:24] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:24] [INFO]   |-- model provider: 'nvidia'

[03:29:24] [INFO]   |-- inference parameters:

[03:29:24] [INFO]   |  |-- generation_type=chat-completion

[03:29:24] [INFO]   |  |-- max_parallel_requests=4

[03:29:24] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:24] [INFO]   |  |-- temperature=1.00

[03:29:24] [INFO]   |  |-- top_p=1.00

[03:29:24] [INFO]   |  |-- max_tokens=2048

[03:29:24] [INFO] ⚡️ Processing llm-text column 'review_summary' with 4 concurrent workers

[03:29:24] [INFO] ⏱️ llm-text column 'review_summary' will report progress after each record

[03:29:24] [INFO]   |-- 🌗 llm-text column 'review_summary' progress: 1/2 (50%) complete, 1 ok, 0 failed, 1.48 rec/s, eta 0.7s

[03:29:24] [INFO]   |-- 🌕 llm-text column 'review_summary' progress: 2/2 (100%) complete, 2 ok, 0 failed, 2.47 rec/s, eta 0.0s

[03:29:25] [INFO] 📊 Model usage summary:

[03:29:25] [INFO]   |-- model: nvidia/nemotron-3-nano-30b-a3b

[03:29:25] [INFO]   |-- tokens: input=1748, output=712, total=2460, tps=463

[03:29:25] [INFO]   |-- requests: success=6, failed=0, total=6, rpm=67

[03:29:25] [INFO] 🙈 Dropping columns: ['customer']

[03:29:25] [INFO] 📐 Measuring dataset column statistics:

[03:29:25] [INFO]   |-- 🎲 column: 'product_category'

[03:29:25] [INFO]   |-- 🎲 column: 'product_subcategory'

[03:29:25] [INFO]   |-- 🎲 column: 'target_age_range'

[03:29:25] [INFO]   |-- 🎲 column: 'review_style'

[03:29:25] [INFO]   |-- 🧩 column: 'customer_name'

[03:29:25] [INFO]   |-- 🧩 column: 'customer_age'

[03:29:25] [INFO]   |-- 🗂️ column: 'product'

[03:29:25] [INFO]   |-- 🗂️ column: 'customer_review'

[03:29:25] [INFO]   |-- 📝 column: 'complaint_analysis'

[03:29:25] [INFO]   |-- 📝 column: 'action_items'

[03:29:25] [INFO]   |-- 📝 column: 'review_summary'

[03:29:25] [INFO] 🥳 Preview complete!

In [12]:

# Run this cell multiple times to cycle through the 2 preview records.
# Look for rows where complaint_analysis and action_items are None (skipped)
# vs rows where they were generated (low-rated reviews).
preview.display_sample_record()

# Run this cell multiple times to cycle through the 2 preview records. # Look for rows where complaint_analysis and action_items are None (skipped) # vs rows where they were generated (low-rated reviews). preview.display_sample_record()

[index: 0]

                                                                                                              
                                              Generated Columns                                               
┏━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ Name                ┃ Value                                                                                ┃
┡━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ product_category    │ Home Office                                                                          │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ product_subcategory │ Storage                                                                              │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ target_age_range    │ 50-65                                                                                │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ review_style        │ detailed                                                                             │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ complaint_analysis  │ None                                                                                 │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ action_items        │ None                                                                                 │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ review_summary      │ A highly praised, sturdy laminated storage cabinet with smooth full‑extension        │
│                     │ drawers and premium metal hardware that fits neatly under desks, blends modern       │
│                     │ aesthetics with safety‑focused design, and delivers reliable, organized performance. │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ product             │ {                                                                                    │
│                     │     'name': 'Expandable Laminated Storage Cabinet',                                  │
│                     │     'description': 'A sturdy, roll-out laminated cabinet designed to fit under desks │
│                     │ or beside office chairs. Features smooth gliding drawers, sturdy metal hardware, and │
│                     │ a sleek, low-profile finish that complements modern home offices. Ideal for          │
│                     │ organizing papers, office supplies, and personal items within easy reach.',          │
│                     │     'price': 84.99                                                                   │
│                     │ }                                                                                    │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ customer_review     │ {                                                                                    │
│                     │     'rating': 5,                                                                     │
│                     │     'customer_mood': 'happy',                                                        │
│                     │     'review': 'The Expandable Laminated Storage Cabinet offers an impressively       │
│                     │ sturdy construction with smooth-gliding, full‑extension drawers that accommodate     │
│                     │ papers, office supplies, and personal items with ease. The low‑profile, laminated    │
│                     │ finish integrates seamlessly with modern home‑office aesthetics, and the robust      │
│                     │ metal hardware conveys a premium feel while maintaining durability. Its compact      │
│                     │ dimensions allow effortless placement beneath desks or beside chairs, maximizing     │
│                     │ work‑space efficiency without compromising accessibility. The streamlined design     │
│                     │ eliminates cabinet handles, reducing snagging risks and enhancing safety when        │
│                     │ positioning furniture. Overall, this storage solution combines functional elegance   │
│                     │ with reliable performance, making it a highly suitable addition to any organized     │
│                     │ workspace.'                                                                          │
│                     │ }                                                                                    │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ customer_name       │ Carl Davis                                                                           │
├─────────────────────┼──────────────────────────────────────────────────────────────────────────────────────┤
│ customer_age        │ 96                                                                                   │
└─────────────────────┴──────────────────────────────────────────────────────────────────────────────────────┘
                                                                                                              

In [13]:

# The preview dataset is available as a pandas DataFrame.
# Notice that complaint_analysis, action_items, and review_summary columns
# reflect the skip behavior: None for skipped rows, generated text otherwise.
preview.dataset

# The preview dataset is available as a pandas DataFrame. # Notice that complaint_analysis, action_items, and review_summary columns # reflect the skip behavior: None for skipped rows, generated text otherwise. preview.dataset

Out[13]:

	product_category	product_subcategory	target_age_range	review_style	customer_name	customer_age	product	customer_review	complaint_analysis	action_items	review_summary
0	Home Office	Storage	50-65	detailed	Carl Davis	96	{'name': 'Expandable Laminated Storage Cabinet...	{'rating': 5, 'customer_mood': 'happy', 'revie...	None	None	A highly praised, sturdy laminated storage cab...
1	Books	Classics	50-65	detailed	Jessica Pacheco	37	{'name': 'Timeless Classics: A Curated Library...	{'rating': 5, 'customer_mood': 'happy', 'revie...	None	None	The Timeless Classics anthology earns a perfec...

📊 Analyze the generated data¶

• Data Designer automatically generates a basic statistical analysis of the generated data.

• This analysis is available via the analysis property of generation result objects.

In [14]:

# Print the analysis as a table.
preview.analysis.to_report()

# Print the analysis as a table. preview.analysis.to_report()

──────────────────────────────────────── 🎨 Data Designer Dataset Profile ─────────────────────────────────────────

                                                                                                                   
                                                 Dataset Overview                                                  
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ number of records               ┃ number of columns               ┃ percent complete records                    ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ 2                               │ 11                              │ 100.0%                                      │
└─────────────────────────────────┴─────────────────────────────────┴─────────────────────────────────────────────┘
                                                                                                                   
                                                                                                                   
                                                🎲 Sampler Columns                                                 
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┓
┃ column name                      ┃        data type ┃               number unique values ┃         sampler type ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━┩
│ product_category                 │           string │                         2 (100.0%) │             category │
├──────────────────────────────────┼──────────────────┼────────────────────────────────────┼──────────────────────┤
│ product_subcategory              │           string │                         2 (100.0%) │          subcategory │
├──────────────────────────────────┼──────────────────┼────────────────────────────────────┼──────────────────────┤
│ target_age_range                 │           string │                          1 (50.0%) │             category │
├──────────────────────────────────┼──────────────────┼────────────────────────────────────┼──────────────────────┤
│ review_style                     │           string │                          1 (50.0%) │             category │
└──────────────────────────────────┴──────────────────┴────────────────────────────────────┴──────────────────────┘
                                                                                                                   
                                                                                                                   
                                                📝 LLM-Text Columns                                                
┏━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┓
┃                         ┃              ┃                            ┃      prompt tokens ┃    completion tokens ┃
┃ column name             ┃    data type ┃       number unique values ┃         per record ┃           per record ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━┩
│ complaint_analysis      │         None │                   0 (0.0%) │     255.5 +/- 69.5 │          1.0 +/- 0.0 │
├─────────────────────────┼──────────────┼────────────────────────────┼────────────────────┼──────────────────────┤
│ action_items            │         None │                   0 (0.0%) │       22.0 +/- 0.0 │          1.0 +/- 0.0 │
├─────────────────────────┼──────────────┼────────────────────────────┼────────────────────┼──────────────────────┤
│ review_summary          │       string │                 2 (100.0%) │     229.0 +/- 70.0 │        53.0 +/- 17.0 │
└─────────────────────────┴──────────────┴────────────────────────────┴────────────────────┴──────────────────────┘
                                                                                                                   
                                                                                                                   
                                             🗂️ LLM-Structured Columns                                             
┏━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┓
┃                      ┃               ┃                            ┃       prompt tokens ┃     completion tokens ┃
┃ column name          ┃     data type ┃       number unique values ┃          per record ┃            per record ┃
┡━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━┩
│ product              │          dict │                 2 (100.0%) │       264.0 +/- 0.0 │         54.5 +/- 37.5 │
├──────────────────────┼───────────────┼────────────────────────────┼─────────────────────┼───────────────────────┤
│ customer_review      │          dict │                 2 (100.0%) │      310.0 +/- 25.0 │        216.0 +/- 96.2 │
└──────────────────────┴───────────────┴────────────────────────────┴─────────────────────┴───────────────────────┘
                                                                                                                   
                                                                                                                   
                                               🧩 Expression Columns                                               
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ column name                       ┃                data type ┃                             number unique values ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ customer_name                     │                   string │                                       2 (100.0%) │
├───────────────────────────────────┼──────────────────────────┼──────────────────────────────────────────────────┤
│ customer_age                      │                   string │                                       2 (100.0%) │
└───────────────────────────────────┴──────────────────────────┴──────────────────────────────────────────────────┘
                                                                                                                   
                                                                                                                   
╭────────────────────────────────────────────────── Table Notes ──────────────────────────────────────────────────╮
│                                                                                                                 │
│  1. All token statistics are based on a sample of max(1000, len(dataset)) records.                              │
│  2. Tokens are calculated using tiktoken's cl100k_base tokenizer.                                               │
│                                                                                                                 │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
                                                                                                                   
───────────────────────────────────────────────────────────────────────────────────────────────────────────────────

🆙 Scale up!¶

• Happy with your preview data?

• Use the create method to submit larger Data Designer generation jobs.

In [15]:

results = data_designer.create(config_builder, num_records=10, dataset_name="tutorial-2")

results = data_designer.create(config_builder, num_records=10, dataset_name="tutorial-2")

[03:29:25] [INFO] 🎨 Creating Data Designer dataset

[03:29:25] [INFO]   |-- 🔒 Jinja rendering engine: secure

[03:29:25] [INFO] ✅ Validation passed

[03:29:25] [INFO] ⛓️ Sorting column configs into a Directed Acyclic Graph

[03:29:25] [INFO] 🩺 Running health checks for models...

[03:29:25] [INFO]   |-- 👀 Checking 'nvidia/nemotron-3-nano-30b-a3b' in provider named 'nvidia' for model alias 'nemotron-nano-v3'...

[03:29:25] [INFO]   |-- ✅ Passed!

[03:29:25] [INFO] ⏳ Processing batch 1 of 1

[03:29:25] [INFO] 🎲 Preparing samplers to generate 10 records across 5 columns

[03:29:25] [INFO] 🧩 Generating column `customer_name` from expression

[03:29:25] [INFO] 🧩 Generating column `customer_age` from expression

[03:29:25] [INFO] 🗂️ llm-structured model config for column 'product'

[03:29:25] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:25] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:25] [INFO]   |-- model provider: 'nvidia'

[03:29:25] [INFO]   |-- inference parameters:

[03:29:25] [INFO]   |  |-- generation_type=chat-completion

[03:29:25] [INFO]   |  |-- max_parallel_requests=4

[03:29:25] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:25] [INFO]   |  |-- temperature=1.00

[03:29:25] [INFO]   |  |-- top_p=1.00

[03:29:25] [INFO]   |  |-- max_tokens=2048

[03:29:25] [INFO] ⚡️ Processing llm-structured column 'product' with 4 concurrent workers

[03:29:25] [INFO] ⏱️ llm-structured column 'product' will report progress after each record

[03:29:26] [INFO]   |-- 🚶 llm-structured column 'product' progress: 1/10 (10%) complete, 1 ok, 0 failed, 1.09 rec/s, eta 8.3s

[03:29:27] [INFO]   |-- 🚶 llm-structured column 'product' progress: 2/10 (20%) complete, 2 ok, 0 failed, 1.73 rec/s, eta 4.6s

[03:29:27] [INFO]   |-- 🐴 llm-structured column 'product' progress: 3/10 (30%) complete, 3 ok, 0 failed, 2.31 rec/s, eta 3.0s

[03:29:27] [INFO]   |-- 🐴 llm-structured column 'product' progress: 4/10 (40%) complete, 4 ok, 0 failed, 2.42 rec/s, eta 2.5s

[03:29:27] [INFO]   |-- 🚗 llm-structured column 'product' progress: 5/10 (50%) complete, 5 ok, 0 failed, 2.54 rec/s, eta 2.0s

[03:29:28] [INFO]   |-- 🚗 llm-structured column 'product' progress: 6/10 (60%) complete, 6 ok, 0 failed, 2.42 rec/s, eta 1.7s

[03:29:28] [INFO]   |-- 🚗 llm-structured column 'product' progress: 7/10 (70%) complete, 7 ok, 0 failed, 2.46 rec/s, eta 1.2s

[03:29:28] [INFO]   |-- ✈️ llm-structured column 'product' progress: 8/10 (80%) complete, 8 ok, 0 failed, 2.63 rec/s, eta 0.8s

[03:29:28] [INFO]   |-- ✈️ llm-structured column 'product' progress: 9/10 (90%) complete, 9 ok, 0 failed, 2.96 rec/s, eta 0.3s

[03:29:29] [INFO]   |-- 🚀 llm-structured column 'product' progress: 10/10 (100%) complete, 10 ok, 0 failed, 2.81 rec/s, eta 0.0s

[03:29:29] [INFO] 🗂️ llm-structured model config for column 'customer_review'

[03:29:29] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:29] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:29] [INFO]   |-- model provider: 'nvidia'

[03:29:29] [INFO]   |-- inference parameters:

[03:29:29] [INFO]   |  |-- generation_type=chat-completion

[03:29:29] [INFO]   |  |-- max_parallel_requests=4

[03:29:29] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:29] [INFO]   |  |-- temperature=1.00

[03:29:29] [INFO]   |  |-- top_p=1.00

[03:29:29] [INFO]   |  |-- max_tokens=2048

[03:29:29] [INFO] ⚡️ Processing llm-structured column 'customer_review' with 4 concurrent workers

[03:29:29] [INFO] ⏱️ llm-structured column 'customer_review' will report progress after each record

[03:29:30] [INFO]   |-- 🌑 llm-structured column 'customer_review' progress: 1/10 (10%) complete, 1 ok, 0 failed, 0.86 rec/s, eta 10.4s

[03:29:31] [INFO]   |-- 🌑 llm-structured column 'customer_review' progress: 2/10 (20%) complete, 2 ok, 0 failed, 1.14 rec/s, eta 7.0s

[03:29:31] [INFO]   |-- 🌘 llm-structured column 'customer_review' progress: 3/10 (30%) complete, 3 ok, 0 failed, 1.18 rec/s, eta 6.0s

[03:29:32] [INFO]   |-- 🌘 llm-structured column 'customer_review' progress: 4/10 (40%) complete, 4 ok, 0 failed, 1.17 rec/s, eta 5.1s

[03:29:33] [INFO]   |-- 🌗 llm-structured column 'customer_review' progress: 5/10 (50%) complete, 5 ok, 0 failed, 1.24 rec/s, eta 4.0s

[03:29:34] [INFO]   |-- 🌗 llm-structured column 'customer_review' progress: 6/10 (60%) complete, 6 ok, 0 failed, 1.27 rec/s, eta 3.1s

[03:29:34] [INFO]   |-- 🌗 llm-structured column 'customer_review' progress: 7/10 (70%) complete, 7 ok, 0 failed, 1.46 rec/s, eta 2.1s

[03:29:34] [INFO]   |-- 🌖 llm-structured column 'customer_review' progress: 8/10 (80%) complete, 8 ok, 0 failed, 1.53 rec/s, eta 1.3s

[03:29:35] [INFO]   |-- 🌖 llm-structured column 'customer_review' progress: 9/10 (90%) complete, 9 ok, 0 failed, 1.46 rec/s, eta 0.7s

[03:29:36] [INFO]   |-- 🌕 llm-structured column 'customer_review' progress: 10/10 (100%) complete, 10 ok, 0 failed, 1.41 rec/s, eta 0.0s

[03:29:36] [INFO] 📝 llm-text model config for column 'complaint_analysis'

[03:29:36] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:36] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:36] [INFO]   |-- model provider: 'nvidia'

[03:29:36] [INFO]   |-- inference parameters:

[03:29:36] [INFO]   |  |-- generation_type=chat-completion

[03:29:36] [INFO]   |  |-- max_parallel_requests=4

[03:29:36] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:36] [INFO]   |  |-- temperature=1.00

[03:29:36] [INFO]   |  |-- top_p=1.00

[03:29:36] [INFO]   |  |-- max_tokens=2048

[03:29:36] [INFO] ⚡️ Processing llm-text column 'complaint_analysis' with 4 concurrent workers

[03:29:36] [INFO] ⏱️ llm-text column 'complaint_analysis' will report progress after each record

[03:29:36] [INFO]   |-- 🌧️ llm-text column 'complaint_analysis' progress: 1/10 (10%) complete, 0 ok, 0 failed, 1 skipped, 984.40 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🌧️ llm-text column 'complaint_analysis' progress: 2/10 (20%) complete, 0 ok, 0 failed, 2 skipped, 1278.36 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🌦️ llm-text column 'complaint_analysis' progress: 3/10 (30%) complete, 0 ok, 0 failed, 3 skipped, 1359.76 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🌦️ llm-text column 'complaint_analysis' progress: 4/10 (40%) complete, 0 ok, 0 failed, 4 skipped, 1489.35 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- ⛅ llm-text column 'complaint_analysis' progress: 5/10 (50%) complete, 0 ok, 0 failed, 5 skipped, 1539.71 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- ⛅ llm-text column 'complaint_analysis' progress: 6/10 (60%) complete, 0 ok, 0 failed, 6 skipped, 1582.19 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- ⛅ llm-text column 'complaint_analysis' progress: 7/10 (70%) complete, 0 ok, 0 failed, 7 skipped, 1586.89 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🌤️ llm-text column 'complaint_analysis' progress: 8/10 (80%) complete, 0 ok, 0 failed, 8 skipped, 1604.89 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🌤️ llm-text column 'complaint_analysis' progress: 9/10 (90%) complete, 0 ok, 0 failed, 9 skipped, 1640.46 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- ☀️ llm-text column 'complaint_analysis' progress: 10/10 (100%) complete, 0 ok, 0 failed, 10 skipped, 1644.11 rec/s, eta 0.0s

[03:29:36] [INFO] 📝 llm-text model config for column 'action_items'

[03:29:36] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:36] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:36] [INFO]   |-- model provider: 'nvidia'

[03:29:36] [INFO]   |-- inference parameters:

[03:29:36] [INFO]   |  |-- generation_type=chat-completion

[03:29:36] [INFO]   |  |-- max_parallel_requests=4

[03:29:36] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:36] [INFO]   |  |-- temperature=1.00

[03:29:36] [INFO]   |  |-- top_p=1.00

[03:29:36] [INFO]   |  |-- max_tokens=2048

[03:29:36] [INFO] ⚡️ Processing llm-text column 'action_items' with 4 concurrent workers

[03:29:36] [INFO] ⏱️ llm-text column 'action_items' will report progress after each record

[03:29:36] [INFO]   |-- 🥚 llm-text column 'action_items' progress: 1/10 (10%) complete, 0 ok, 0 failed, 1 skipped, 1360.69 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🥚 llm-text column 'action_items' progress: 2/10 (20%) complete, 0 ok, 0 failed, 2 skipped, 1817.71 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐣 llm-text column 'action_items' progress: 3/10 (30%) complete, 0 ok, 0 failed, 3 skipped, 2048.70 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐣 llm-text column 'action_items' progress: 4/10 (40%) complete, 0 ok, 0 failed, 4 skipped, 2181.22 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐥 llm-text column 'action_items' progress: 5/10 (50%) complete, 0 ok, 0 failed, 5 skipped, 2268.57 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐥 llm-text column 'action_items' progress: 6/10 (60%) complete, 0 ok, 0 failed, 6 skipped, 2351.54 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐥 llm-text column 'action_items' progress: 7/10 (70%) complete, 0 ok, 0 failed, 7 skipped, 2408.85 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐤 llm-text column 'action_items' progress: 8/10 (80%) complete, 0 ok, 0 failed, 8 skipped, 2467.52 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐤 llm-text column 'action_items' progress: 9/10 (90%) complete, 0 ok, 0 failed, 9 skipped, 2478.11 rec/s, eta 0.0s

[03:29:36] [INFO]   |-- 🐔 llm-text column 'action_items' progress: 10/10 (100%) complete, 0 ok, 0 failed, 10 skipped, 2494.27 rec/s, eta 0.0s

[03:29:36] [INFO] 📝 llm-text model config for column 'review_summary'

[03:29:36] [INFO]   |-- model: 'nvidia/nemotron-3-nano-30b-a3b'

[03:29:36] [INFO]   |-- model alias: 'nemotron-nano-v3'

[03:29:36] [INFO]   |-- model provider: 'nvidia'

[03:29:36] [INFO]   |-- inference parameters:

[03:29:36] [INFO]   |  |-- generation_type=chat-completion

[03:29:36] [INFO]   |  |-- max_parallel_requests=4

[03:29:36] [INFO]   |  |-- extra_body={'chat_template_kwargs': {'enable_thinking': False}}

[03:29:36] [INFO]   |  |-- temperature=1.00

[03:29:36] [INFO]   |  |-- top_p=1.00

[03:29:36] [INFO]   |  |-- max_tokens=2048

[03:29:36] [INFO] ⚡️ Processing llm-text column 'review_summary' with 4 concurrent workers

[03:29:36] [INFO] ⏱️ llm-text column 'review_summary' will report progress after each record

[03:29:37] [INFO]   |-- 😴 llm-text column 'review_summary' progress: 1/10 (10%) complete, 1 ok, 0 failed, 1.89 rec/s, eta 4.8s

[03:29:37] [INFO]   |-- 😴 llm-text column 'review_summary' progress: 2/10 (20%) complete, 2 ok, 0 failed, 2.76 rec/s, eta 2.9s

[03:29:37] [INFO]   |-- 🥱 llm-text column 'review_summary' progress: 3/10 (30%) complete, 3 ok, 0 failed, 3.69 rec/s, eta 1.9s

[03:29:37] [INFO]   |-- 🥱 llm-text column 'review_summary' progress: 4/10 (40%) complete, 4 ok, 0 failed, 3.99 rec/s, eta 1.5s

[03:29:37] [INFO]   |-- 😐 llm-text column 'review_summary' progress: 5/10 (50%) complete, 5 ok, 0 failed, 3.53 rec/s, eta 1.4s

[03:29:38] [INFO]   |-- 😐 llm-text column 'review_summary' progress: 6/10 (60%) complete, 6 ok, 0 failed, 4.05 rec/s, eta 1.0s

[03:29:38] [INFO]   |-- 😐 llm-text column 'review_summary' progress: 7/10 (70%) complete, 7 ok, 0 failed, 4.41 rec/s, eta 0.7s

[03:29:38] [INFO]   |-- 😊 llm-text column 'review_summary' progress: 8/10 (80%) complete, 8 ok, 0 failed, 4.69 rec/s, eta 0.4s

[03:29:38] [INFO]   |-- 😊 llm-text column 'review_summary' progress: 9/10 (90%) complete, 9 ok, 0 failed, 4.39 rec/s, eta 0.2s

[03:29:38] [INFO]   |-- 🤩 llm-text column 'review_summary' progress: 10/10 (100%) complete, 10 ok, 0 failed, 4.77 rec/s, eta 0.0s

[03:29:38] [INFO] 🙈 Dropping columns: ['customer']

[03:29:38] [INFO] 📊 Model usage summary:

[03:29:38] [INFO]   |-- model: nvidia/nemotron-3-nano-30b-a3b

[03:29:38] [INFO]   |-- tokens: input=9107, output=4035, total=13142, tps=999

[03:29:38] [INFO]   |-- requests: success=30, failed=0, total=30, rpm=136

[03:29:38] [INFO] 📐 Measuring dataset column statistics:

[03:29:38] [INFO]   |-- 🎲 column: 'product_category'

[03:29:38] [INFO]   |-- 🎲 column: 'product_subcategory'

[03:29:38] [INFO]   |-- 🎲 column: 'target_age_range'

[03:29:38] [INFO]   |-- 🎲 column: 'review_style'

[03:29:38] [INFO]   |-- 🧩 column: 'customer_name'

[03:29:38] [INFO]   |-- 🧩 column: 'customer_age'

[03:29:38] [INFO]   |-- 🗂️ column: 'product'

[03:29:38] [INFO]   |-- 🗂️ column: 'customer_review'

[03:29:39] [INFO]   |-- 📝 column: 'complaint_analysis'

[03:29:39] [INFO]   |-- 📝 column: 'action_items'

[03:29:39] [INFO]   |-- 📝 column: 'review_summary'

In [16]:

# Load the generated dataset as a pandas DataFrame.
dataset = results.load_dataset()

dataset.head()

# Load the generated dataset as a pandas DataFrame. dataset = results.load_dataset() dataset.head()

Out[16]:

	product_category	product_subcategory	target_age_range	review_style	customer_name	customer_age	product	customer_review	complaint_analysis	action_items	review_summary
0	Home Office	Lighting	65+	structured with bullet points	Alexis Lowe	44	{'description': 'A user-friendly, height-adjus...	{'customer_mood': 'happy', 'rating': 5, 'revie...	None	None	The SunGlow Adjustable Warm LED Desk Lamp offe...
1	Home & Kitchen	Organization	25-35	detailed	David Jensen	88	{'description': 'A sleek, stackable, magnetic ...	{'customer_mood': 'happy', 'rating': 4, 'revie...	None	None	The Modular Kitchen Organizer impressed this r...
2	Electronics	Headphones	18-25	rambling	Austin Phillips	69	{'description': 'Compact, ultra-lightweight tr...	{'customer_mood': 'happy', 'rating': 4, 'revie...	None	None	Despite its pocket‑thin design and vibrant RGB...
3	Home Office	Office Supplies	50-65	brief	Bonnie Nelson	19	{'description': 'A premium, fully adjustable e...	{'customer_mood': 'happy', 'rating': 5, 'revie...	None	None	The Elegant Ergonomic Executive Chair impresse...
4	Home Office	Desks	18-25	rambling	Veronica Bailey	90	{'description': 'Ergonomic standing desk with ...	{'customer_mood': 'happy', 'rating': 4, 'revie...	None	None	The AeroDesk Pro impresses with its sleek desi...

In [17]:

# Load the analysis results into memory.
analysis = results.load_analysis()

analysis.to_report()

# Load the analysis results into memory. analysis = results.load_analysis() analysis.to_report()

──────────────────────────────────────── 🎨 Data Designer Dataset Profile ─────────────────────────────────────────

                                                                                                                   
                                                 Dataset Overview                                                  
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ number of records               ┃ number of columns               ┃ percent complete records                    ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ 10                              │ 11                              │ 100.0%                                      │
└─────────────────────────────────┴─────────────────────────────────┴─────────────────────────────────────────────┘
                                                                                                                   
                                                                                                                   
                                                🎲 Sampler Columns                                                 
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┓
┃ column name                      ┃        data type ┃               number unique values ┃         sampler type ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━┩
│ product_category                 │           string │                          4 (40.0%) │             category │
├──────────────────────────────────┼──────────────────┼────────────────────────────────────┼──────────────────────┤
│ product_subcategory              │           string │                        10 (100.0%) │          subcategory │
├──────────────────────────────────┼──────────────────┼────────────────────────────────────┼──────────────────────┤
│ target_age_range                 │           string │                          5 (50.0%) │             category │
├──────────────────────────────────┼──────────────────┼────────────────────────────────────┼──────────────────────┤
│ review_style                     │           string │                          4 (40.0%) │             category │
└──────────────────────────────────┴──────────────────┴────────────────────────────────────┴──────────────────────┘
                                                                                                                   
                                                                                                                   
                                                📝 LLM-Text Columns                                                
┏━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━┓
┃                         ┃              ┃                            ┃      prompt tokens ┃    completion tokens ┃
┃ column name             ┃    data type ┃       number unique values ┃         per record ┃           per record ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━┩
│ complaint_analysis      │         None │                   0 (0.0%) │     266.5 +/- 82.3 │          1.0 +/- 0.0 │
├─────────────────────────┼──────────────┼────────────────────────────┼────────────────────┼──────────────────────┤
│ action_items            │         None │                   0 (0.0%) │       22.0 +/- 0.0 │          1.0 +/- 0.0 │
├─────────────────────────┼──────────────┼────────────────────────────┼────────────────────┼──────────────────────┤
│ review_summary          │       string │                10 (100.0%) │     238.0 +/- 81.9 │        57.5 +/- 10.9 │
└─────────────────────────┴──────────────┴────────────────────────────┴────────────────────┴──────────────────────┘
                                                                                                                   
                                                                                                                   
                                             🗂️ LLM-Structured Columns                                             
┏━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━┓
┃                      ┃               ┃                            ┃       prompt tokens ┃     completion tokens ┃
┃ column name          ┃     data type ┃       number unique values ┃          per record ┃            per record ┃
┡━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━┩
│ product              │          dict │                10 (100.0%) │       265.0 +/- 0.5 │         77.5 +/- 16.6 │
├──────────────────────┼───────────────┼────────────────────────────┼─────────────────────┼───────────────────────┤
│ customer_review      │          dict │                10 (100.0%) │      332.0 +/- 15.3 │        230.5 +/- 86.4 │
└──────────────────────┴───────────────┴────────────────────────────┴─────────────────────┴───────────────────────┘
                                                                                                                   
                                                                                                                   
                                               🧩 Expression Columns                                               
┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ column name                       ┃                data type ┃                             number unique values ┃
┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ customer_name                     │                   string │                                      10 (100.0%) │
├───────────────────────────────────┼──────────────────────────┼──────────────────────────────────────────────────┤
│ customer_age                      │                   string │                                      10 (100.0%) │
└───────────────────────────────────┴──────────────────────────┴──────────────────────────────────────────────────┘
                                                                                                                   
                                                                                                                   
╭────────────────────────────────────────────────── Table Notes ──────────────────────────────────────────────────╮
│                                                                                                                 │
│  1. All token statistics are based on a sample of max(1000, len(dataset)) records.                              │
│  2. Tokens are calculated using tiktoken's cl100k_base tokenizer.                                               │
│                                                                                                                 │
╰─────────────────────────────────────────────────────────────────────────────────────────────────────────────────╯
                                                                                                                   
───────────────────────────────────────────────────────────────────────────────────────────────────────────────────

⏭️ Next Steps¶

Check out the following notebook to learn more about:

• Seeding synthetic data generation with an external dataset

• Providing images as context

• Generating images