Quickstart: Gen AI evaluation workflow
This page shows you how to perform a model-based evaluation with Gen AI Evaluation Service using the Vertex AI SDK for Python.
Before you begin
-
Sign in to your Google Cloud account. If you're new to Google Cloud, create an account to evaluate how our products perform in real-world scenarios. New customers also get $300 in free credits to run, test, and deploy workloads. In the Google Cloud console, on the project selector page, select or create a Google Cloud project.
Make sure that billing is enabled for your Google Cloud project.
In the Google Cloud console, on the project selector page, select or create a Google Cloud project.
Make sure that billing is enabled for your Google Cloud project.
Install the Vertex AI SDK for Python with Gen AI Evaluation Service dependency:
!pip install google-cloud-aiplatform[evaluation]
Set up your credentials. If you are running this quickstart in Colaboratory, run the following:
from google.colab import auth auth.authenticate_user()
For other environments, refer to Authenticate to Vertex AI.
Import libraries
Import your libraries and set up your project and location.
import pandas as pd
import vertexai
from vertexai.evaluation import EvalTask, PointwiseMetric, PointwiseMetricPromptTemplate
PROJECT_ID = "PROJECT_ID"
LOCATION = "LOCATION"
EXPERIMENT_NAME = "EXPERIMENT_NAME"
vertexai.init(
project=PROJECT_ID,
location=LOCATION,
)
Set up evaluation metrics based on your criteria
The following metric definition evaluates the text quality generated from a large language model based on two criteria: Fluency
and Entertaining
. The code defines a metric called custom_text_quality
using those two criteria:
custom_text_quality = PointwiseMetric(
metric="custom_text_quality",
metric_prompt_template=PointwiseMetricPromptTemplate(
criteria={
"fluency": (
"Sentences flow smoothly and are easy to read, avoiding awkward"
" phrasing or run-on sentences. Ideas and sentences connect"
" logically, using transitions effectively where needed."
),
"entertaining": (
"Short, amusing text that incorporates emojis, exclamations and"
" questions to convey quick and spontaneous communication and"
" diversion."
),
},
rating_rubric={
"1": "The response performs well on both criteria.",
"0": "The response is somewhat aligned with both criteria",
"-1": "The response falls short on both criteria",
},
),
)
Prepare your dataset
Add the following code to prepare your dataset:
responses = [
# An example of good custom_text_quality
"Life is a rollercoaster, full of ups and downs, but it's the thrill that keeps us coming back for more!",
# An example of medium custom_text_quality
"The weather is nice today, not too hot, not too cold.",
# An example of poor custom_text_quality
"The weather is, you know, whatever.",
]
eval_dataset = pd.DataFrame({
"response" : responses,
})
Run evaluation with your dataset
Run the evaluation:
eval_task = EvalTask(
dataset=eval_dataset,
metrics=[custom_text_quality],
experiment=EXPERIMENT_NAME
)
pointwise_result = eval_task.evaluate()
View the evaluation results for each response in the metrics_table
:
pointwise_result.metrics_table
Clean up
To avoid incurring charges to your Google Cloud account for the resources used on this page, follow these steps.
Delete the ExperimentRun
created by the evaluation:
vertexai.ExperimentRun(
run_name=pointwise_result.metadata["experiment_run"],
experiment=pointwise_result.metadata["experiment"],
).delete()