Welcome and thank you for joining the Mantium team for our upcoming workshop! To prepare you for participating, we have included a few steps below to help you get set up for success.
Note: To participate in this demo, you are only required to have a Mantium account username, password, and prompt_id for a roBERTa prompt.
For creating other prompts, you will need a Mantium account and an API key for an LLM (large language model) provider - this short Loom video will take you through the process. If you do not wish to use a non-Mantium LLM provider, you are welcome to use Mantium GPT-J without an API key!
Join our Discord server and share your ideas, teach, and learn about the world of growing AI technologies. New to coding? No problem. This is an inclusive space!
To get an understanding of the type of application you will be building in the workshop, we’ve included a link to an interactive demo prompt. This is the Mantium Video Game Generator - all you have to do is type in a creative video game title for a game that might not yet exist, and the language model will generate a unique description! Try it out as many times as you’d like, and feel free to share any fun responses with us.
The Mantium developer hub can be found here. If you’d like to learn more, this is where you’ll find tutorials and documentation!
To build your own sentiment analysis prompt, create one in the Mantium UI.
Step 1: Navigate to AI Manager > Prompts
Step 2: Click Add new prompt
Step 3: Choose Mantium as a Provider and Roberta_Base_Goemotions_Classification as the endpoint.
Step 4: Fill in the Name of Prompt and click Save
Step 5: To locate the prompt_id value (used in the next step), select your prompt from the Prompts menu and copy the hash value found in the Endpoint API property.
Reminder: please never use unencrypted environment variables in production, and never push them to a public repo!
Step 1: Update .env-demo to .env
$ mv .env-demo .env
Step 2: Replace placeholder strings in .env with your Mantium credentials
Step 1: Change into a directory where you keep your projects
$ cd ~/path/to/your/projects
Step 2: Clone this repository into your projects directory
$ git clone url/to/this/repository
Step 3: Change into the directory where you cloned the repository
$ cd ~/path/to/your/projects/lwt_summit
Step 4: Create a virtual environment
$ mkvirtualenv mantium_python_project
Step 7: Install the project dependencies using pip
$ pip install -r requirements.txt
If you use other methods, feel free to adapt these steps to your needs!
$ export PYTHONPATH=./app/:./tests/
With FastAPI and Uvicorn you should be able to run the backend with one simple command:
$ uvicorn main:app --reload
To send requests:
Step 1: Navigate to http://127.0.0.1:8000/docs in your browser
Step 2: Expand the POST route by clicking the dropdown arrow on the right and click Try it out
Step 3: Modify the value of "input" with a string that you'd like to have analyzed and click Execute!
$ pytest tests/*