Prompt Evaluator

Automate the process of generating responses to prompts and log into a postgres database.

• 👉 Enter one or more questions. Each subsequent question leverages the conversation history so you can chain questions together.
• 👉 Select one or more models to run the questions through
• 👉 Specify a base_url to retrieve list of models (configured to use litellm and ollama as a openai proxy server and inference engine)
• 👉 Log results to an external postgres db

Features:

• ✅ Up-to-date Dependencies
• ✅ Modern UI
• ✅

Start with Docker

In case the starter was built with Docker support, here is the start up CMD:

$ docker-compose up --build

Once the above command is finished, the new app is started on http://localhost:5085

Manual Build

Download/Clone the sources

$ git clone https://github.com/<THIS_REPO>.git
$ cd <LOCAL_Directory>

Install modules via VENV

$ virtualenv env
$ source env/bin/activate
$ pip install -r requirements.txt

Set Up Database

$ python manage.py makemigrations
$ python manage.py migrate

Generate your API (optional)

$ python manage.py generate-api -f

Start the App

$ python manage.py runserver

At this point, the app runs at http://127.0.0.1:8000/.

---

Starter built with Django App Generator, a free service provided by AppSeed.