Elhusseny/DigitalAhmed-V3-qwen2.5-0.5B

TEXT GENERATIONConcurrency Cost:1Model Size:0.5BQuant:BF16Ctx Length:32kTool Calling:SupportedPublished:Jun 28, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

Elhusseny/DigitalAhmed-V3-qwen2.5-0.5B is a 0.5 billion parameter language model, fine-tuned by Ahmed Hussein using LoRA on a Qwen2.5-0.5B-Instruct base. This model is specifically optimized for generating responses in Egyptian Arabic, embodying the persona of "أحمد الرقمي" (Digital Ahmed). It excels in instruction-following tasks within its specialized cultural and linguistic context.

Loading preview...

DigitalAhmed-V3 (Qwen2.5-0.5B) Overview

This model, developed by Ahmed Hussein, is a specialized 0.5 billion parameter language model built upon the Qwen2.5-0.5B-Instruct base. It has been fine-tuned using the LoRA method via Unsloth, leveraging a custom Egyptian Arabic instruction dataset. The primary goal of DigitalAhmed-V3 is to replicate the persona of "أحمد الرقمي" (Digital Ahmed) and generate culturally relevant responses in Egyptian Arabic.

Key Capabilities

  • Egyptian Arabic Generation: Optimized for producing text in the Egyptian Arabic dialect.
  • Persona Emulation: Designed to embody the characteristics and style of the "Digital Ahmed" persona.
  • Instruction Following: Capable of responding to instructions within its specialized linguistic domain.

Model Variants

DigitalAhmed-V3 is available in different quantization levels, including an F16 variant (1GB) and a Q8_0 variant (0.5GB), offering flexibility for deployment based on performance and resource constraints.

Good For

  • Applications requiring text generation in Egyptian Arabic.
  • Projects aiming to integrate a specific cultural persona into conversational AI.
  • Research and development focused on dialect-specific language models.