uabali/gemma4-e4b-TR
VISIONConcurrent Unit Cost:1Model Size:7.9BQuant:FP8Context Size:32kTool Calling:SupportedPublished:Apr 17, 2026License:apache-2.0Architecture:Transformer0.0K Open Weights Featherless Exclusive Cold
The uabali/gemma4-e4b-TR model is a 7.9 billion parameter Gemma 4 E4B variant, fine-tuned by uabali using LoRA (QLoRA) on the Turkish RAG dataset. Optimized for Retrieval-Augmented Generation (RAG) tasks in Turkish, this GGUF-formatted model is designed for high-quality text generation within an 8192-token context length. It specializes in providing accurate and contextually relevant responses for Turkish language applications.
Loading preview...
Model Overview
The uabali/gemma4-e4b-TR is a Turkish RAG-optimized GGUF version of the Gemma 4 E4B model. Developed by uabali, this model was fine-tuned using LoRA (QLoRA) specifically on a Turkish RAG dataset, making it highly specialized for Retrieval-Augmented Generation tasks in the Turkish language.
Key Capabilities
- Turkish RAG Optimization: Fine-tuned on the Metin/WikiRAG-TR dataset to excel in Turkish RAG scenarios.
- GGUF Format: Provided in GGUF format, suitable for efficient deployment and inference with tools like
llama.cpp. - Base Model: Built upon the
google/gemma-4-e4barchitecture. - Context Length: Supports a context length of 8192 tokens, enabling processing of substantial Turkish text inputs for RAG.
Good For
- High-quality Turkish RAG: Ideal for applications requiring accurate and contextually relevant responses from Turkish knowledge bases.
- Turkish language processing: Specialized for tasks involving understanding and generating Turkish text within a RAG framework.
- Efficient deployment: The GGUF format allows for optimized performance on various hardware, particularly with
llama.cpp.