Skip to content

Latest commit

 

History

History
32 lines (24 loc) · 1.55 KB

README.md

File metadata and controls

32 lines (24 loc) · 1.55 KB

AmharicAI-AdGen

Project Goal:

To generate creative and relevant Amharic ad content for Telegram channels, increasing the effectiveness of AiQEM's Adbar advertising solution.

Background:

AiQEM is an Ethiopian startup specializing in AI and Blockchain solutions for businesses. Adbar is their flagship AI-powered Telegram ad solution for targeted ad placement. Growing Telegram use in Ethiopia necessitates adaptation of AI strategies for content generation.

Project Proposal:

Develop an Amharic RAG (Retrieval Augmented Generation) pipeline for automatically creating ad content. Leverage existing open-source LLM models with Amharic capabilities like: Nous Hermes Mistral 8 7B Amharic-finetuned LLama 2 versions (Samuael/llama-2-7b-tebot-amharic or iocuydi/llama-2-amharic-3784m) Further fine-tune the chosen model on Amharic ad text data specific to Telegram channels.

Use the RAG pipeline to generate:

Catchy ad headlines and body text based on campaign briefs (brand, product info) Content tailored to the chosen Telegram channel's theme and audience

Key Success Factors:

High-quality Amharic text embedding and generation capabilities of the chosen LLM model. Fine-tuning on relevant ad text data to ensure content relevance and creativity. Integration with Adbar platform for seamless ad generation and deployment.

Expected Outcomes:

Increased Adbar-generated Amharic ad content led to higher click-through and engagement rates. Improved ad targeting and relevance across the Telegram ecosystem. Improved brand positioning and awareness among AiQEM clients.