---
base_model: ytu-ce-cosmos/Turkish-Llama-8b-DPO-v0.1
tags:
- text-generation-inference
- transformers
- unsloth
- llama
- trl
- reasoning
- bilingual
license: apache-2.0
language:
- tr
- en
new_version: RefinedNeuro/RN_TR_R2
---

<p align="center">
  <img src="https://huggingface.co/RefinedNeuro/RN_TR_R1/resolve/main/rntrr1.png" width="600"/>
</p>

# 🧠 RN_TR_R1 - Turkish-English Reasoning Chat Model (8.03B)

RN_TR_R1 is an open-source, bilingual reasoning chat model fine-tuned on Turkish and English dialogue. It is optimized for instruction-following, multi-step reasoning, and real-time conversation. Built on LLaMA architecture and trained 2x faster with [Unsloth](https://unsloth.ai) + TRL.

## ✨ Highlights

- 🗣️ **Bilingual**: Turkish-first with strong English understanding  
- ⚡ **2x Training Speed**: Thanks to Unsloth + TRL  
- 🧠 **Reasoning Ready**: Handles complex instructions, multi-turn logic, and structured thinking  
- 🧩 **Conversational Tuning**: Optimized for chat + instruction formats  
- 🌐 **Available on**: [Ollama](https://ollama.com/library/rn_tr_r1), [Hugging Face](https://huggingface.co/RefinedNeuro/RN_TR_R1)

## 🚀 Quick Start

```python
# via Ollama
ollama run RefinedNeuro/RN_TR_R1

# or via Transformers
from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("RefinedNeuro/RN_TR_R1")
tokenizer = AutoTokenizer.from_pretrained("RefinedNeuro/RN_TR_R1")