nileagi/swahili-chat-gemma-4-e4b-merged-16bit

VISIONConcurrency Cost:1Model Size:7.9BQuant:FP8Ctx Length:32kTool Calling:SupportedPublished:Jun 17, 2026License:apache-2.0Architecture:Transformer Open Weights Cold

The nileagi/swahili-chat-gemma-4-e4b-merged-16bit model is a 7.9 billion parameter Gemma-4 based language model, finetuned by nileagi. This model was specifically trained using Unsloth and Huggingface's TRL library, achieving 2x faster training. It is optimized for chat-based applications, particularly in Swahili, leveraging its efficient training methodology.

Loading preview...

Model Overview

The nileagi/swahili-chat-gemma-4-e4b-merged-16bit is a 7.9 billion parameter language model, finetuned by nileagi. It is based on the Gemma-4 architecture and was developed using unsloth/gemma-4-e4b-unsloth-bnb-4bit as its base model. The training process leveraged Unsloth and Huggingface's TRL library, which enabled a 2x faster finetuning compared to standard methods.

Key Capabilities

  • Efficient Training: Utilizes Unsloth for significantly faster finetuning.
  • Gemma-4 Architecture: Built upon the robust Gemma-4 base model.
  • Swahili Chat Focus: Optimized for chat-based interactions, likely with a focus on the Swahili language given the model name.

Good For

  • Swahili Language Applications: Ideal for developing chatbots or conversational AI systems requiring Swahili language understanding and generation.
  • Resource-Efficient Deployment: The 16-bit merged format suggests a balance between performance and reduced memory footprint, potentially suitable for deployment on various hardware.
  • Research and Development: Provides a foundation for further experimentation and finetuning on Gemma-4 models with efficient training techniques.