afkfatih/Gemma3-4B-TR

VISIONPricing:Input $0.2 / Output $0.4Concurrent Unit Cost:1Model Size:4.3BQuant:BF16Context Size:32kPublished:Oct 13, 2025Architecture:Transformer Featherless Exclusive Cold

afkfatih/Gemma3-4B-TR is a 4.3 billion parameter language model fine-tuned by afkfatih specifically for Turkish language tasks, based on the Google Gemma 3 4B foundational model. It utilizes LoRA fine-tuning on a 51k-example Turkish dataset, achieving successful performance in Turkish natural language understanding and generation. This model is optimized for applications like Turkish chatbots, Q&A systems, and text generation, offering a specialized solution for Turkish NLP needs.

Loading preview...

Overview

afkfatih/Gemma3-4B-TR is a 4.3 billion parameter language model developed by afkfatih, specifically fine-tuned for Turkish language tasks. It is built upon the Google Gemma 3 4B base model and adapted using LoRA (Low-Rank Adaptation) with a dedicated Turkish dataset containing over 51,000 examples. This model aims to provide strong performance in Turkish NLP applications, even with its current single-epoch training.

Key Capabilities

  • Turkish Language Proficiency: Optimized for understanding and generating text in Turkish.
  • Chatbot and Assistant Development: Suitable for creating conversational AI in Turkish.
  • Question Answering: Can be used for Turkish Q&A systems.
  • Text Generation: Capable of producing various forms of Turkish text.
  • Natural Language Understanding (NLU): Supports NLU tasks in Turkish.

Good for

  • Turkish Chatbot and Assistant Applications: Ideal for building AI agents that interact in Turkish.
  • Turkish Q&A Systems: Developing systems that answer questions in Turkish.
  • Turkish Text Generation: Use cases requiring the creation of Turkish content.
  • Customer Service Automation: Automating interactions with Turkish-speaking customers.
  • Educational Technologies: Applications in Turkish-language learning or content creation.

Finetunes
1 models