- The JokeGPT joke generator will generate jokes leveraging large language models.
- Leveraging NextJS and OpenAI, the JokeGPT generates jokes using various:
- Topics (work, people, animals, food, television, etc)
- Tones (witty, sarcastic, silly, dark, goofy, etc)
- Types (pun, knock-knock, story, etc)
- "Temperature" of it (how much randomness/fun you want to add to the joke)
- Please see JokeGPT-[uniqueID] directories for individual project contributions.
To consolidate the knowledge acquired in this week, the students should complete the following project:
-
Create a github repository for your project
-
Add all members of your group
-
Create a README.md file with the description of your project
-
Create a new application from scratch using NextJS
-
Create a page for generating jokes using AI
-
Add a feature for users to customize the Joke Parameters
- You can pick what kind of parameters you would like to offer for your users
- For example, you could allow users to pick a topic from a list of options (work, people, animals, food, television, etc), a tone for the joke (witty, sarcastic, silly, dark, goofy, etc), the kind of joke (pun, knock-knock, story, etc), and the "temperature" of it (how much randomness/fun you want to add to the joke)
- Consider how you're going to assemble the prompt for the AI model to abide to these parameters
-
After configuring the parameters, users should click a button to generate the joke, and the generated response must be returned to the user's screen inside the same page
-
Add a feature for the AI to evaluate if the generated jokes are funny or not, appropriated or not, offensive or not, and other criteria you might judge important
-
Run the Text Generation WebUI application on your local environment and use it to serve the text generation tasks for your application
-
Submit your project in the submission form
You should find your group in the Discord AI Bootcamp Channel
If you can't find your group, please contact the program manager through Discord or email