AI21Labs/SageMaker

AI21Labs/SageMaker

Description: Examples for using AI21's models through Amazon SageMaker

Language: Jupyter Notebook

Stars: 32

Forks: 16

Open issues: 1

Created: 2022-11-22T14:42:37Z

Pushed: 2024-12-01T12:15:11Z

Default branch: main

Fork: no

Archived: no

README:

Deploy AI21 Labs Language Models on AWS SageMaker

Examples for using Jurassic-2 models through Amazon SageMaker.

Installation

You can install the package using pip:

pip install -U "ai21[AWS]"

And then import the package:

import ai21

Quickstart

In order to use this library, you must configure your AWS credentials.

All set up? Make your first interaction with Jurassic-2:

response = ai21.Completion.execute(destination=ai21.SageMakerDestination("j2-ultra"),
prompt="To be, or",
maxTokens=4,
temperature=0)

print(response['completions'][0]['data']['text'])
# not to be: that is the question

For more comprehensive quickstart guides, see:

Foundation models

[Jurassic-2 Ultra](J2_Ultra_example_model_use.ipynb)

[Jurassic-2 Mid Instruct](J2_Mid_example_model_use.ipynb)

[Jurassic-2 Light](J2_Light_example_model_use.ipynb)

Task-specific models

[AI21 Summarize](AI21_Summarize_example_model_use.ipynb)

[AI21 Contextual Answers](AI21_ContextualAnswers_example_model_use.ipynb)

[AI21 Paraphrase](AI21_Paraphrase_example_model_use.ipynb)

[AI21 Grammatical Error Correction (GEC)](AI21_GEC_example_model_use.ipynb)

Learn more

• AI21 Studio website

• Our technical docs

• AI21 Studio playground - no code environment to test and perfect your prompts