Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/mattsavarino/rai-content-safety

Responsible AI simulations and evaluations using Azure Content Safety
https://github.com/mattsavarino/rai-content-safety

ai automation azure responsible-ai

Last synced: about 1 month ago
JSON representation

Responsible AI simulations and evaluations using Azure Content Safety

Host: GitHub
URL: https://github.com/mattsavarino/rai-content-safety
Owner: mattsavarino
License: mit
Created: 2024-08-22T14:34:26.000Z (5 months ago)
Default Branch: main
Last Pushed: 2024-08-26T22:17:38.000Z (5 months ago)
Last Synced: 2024-10-15T20:41:00.173Z (3 months ago)
Topics: ai, automation, azure, responsible-ai
Language: Jupyter Notebook
Homepage:
Size: 11.7 KB
Stars: 0
Watchers: 1
Forks: 0
Open Issues: 0
Metadata Files:
- Readme: README.md
- License: LICENSE

Awesome Lists containing this project

README

        # Responsible AI: Content Safety

This repo helps automate responsible AI evaluations for content safety.

You can read more details in the [Azure documentation](https://learn.microsoft.com/en-us/azure/ai-studio/how-to/develop/simulator-interaction-data).

## Prerequisites

* Python v3.12+

* Azure access

## ☁️ Azure Setup

1. [Create Azure AI Studio resource](https://portal.azure.com/#create/Microsoft.AzureAIStudio) (note capacity below)

1. Create new project within AI Studio

    **Regional Capacity**


    As of August 2024, the following regions work. [View status](https://learn.microsoft.com/en-us/azure/ai-studio/concepts/evaluation-metrics-built-in?tabs=warning#risk-and-safety-metrics)

    | Region | TPM |

    |---|---|

    | Sweden Central | 450k |

    | France Central | 380k |

    | UK South | 280k |

    | East US 2 | 80k |

## 🐍 Python Setup

1. Clone this repo, open a terminal, and change directory

1. Rename [.env.sample](.env.sample) to `.env`

1. Edit `.env` to add your keys for various Azure services

1. Create virtual environment: `python3 -m venv .venv`

1. Activate virtual environment:

    * Windows: `.\.venv\Scripts\activate`

    * Linux or macOS: `source ./.venv/bin/activate`

1. Install [required packages](./requirements.txt): `pip install -r requirements.txt`

## 🌡️ Run Evaluations

1. Simply open and run [the Python notebooks](./notebooks/).

1. Update `call_endpoint()` function [here](./notebooks/01_content_safety_eval.ipynb).

    * Call your custom API endpoint.

    * Process API response to extract AI text.

3. Monitor the evaluation log and review generated eval files.

4. Generate summary of eval files [here](./notebooks/02_analyze_eval.ipynb).

## ⚙️ Conclusion

* Be responsible with your AI.


* Continuously evaluate for content safety.

* [Learn more about responsible AI](https://www.microsoft.com/en-us/ai/responsible-ai)