wiroai turkish llm 9b

Introduction

WiroAI's WIROAI-TURKISH-LLM-9B is a robust language model designed to support the Turkish language and culture. Built on Google's Gemma architecture, this model excels in Turkish natural language processing tasks through fine-tuning with over 500,000 high-quality Turkish instructions.

Architecture

The model employs a decoder-only transformer architecture, based on Google's Gemma 2 9B. The training process specifically targets Turkish cultural and linguistic nuances to enhance performance in local contexts.

Training

The WIROAI-TURKISH-LLM-9B model is fine-tuned using the LoRA method without quantization. It has been trained on a dataset consisting of more than 500,000 Turkish instructions, ensuring a comprehensive understanding of the language and culture.

Guide: Running Locally

To run the WIROAI-TURKISH-LLM-9B model locally, follow these steps:

1. Install Required Libraries: Ensure you have the Hugging Face Transformers and PyTorch libraries installed.

   pip install transformers torch
   

2. Import the Model: Use the Transformers library to load the model and pipeline.

   import transformers
   import torch
   
   model_id = "WiroAI/wiroai-turkish-llm-9b"
   pipeline = transformers.pipeline(
       "text-generation",
       model=model_id,
       model_kwargs={"torch_dtype": torch.bfloat16},
       device_map="auto",
   )
   

3. Prepare and Run the Model: Set your input instruction and use the pipeline to generate text.

   instruction = "Bana İstanbul ile alakalı bir sosyal medya postu hazırlar mısın?"
   outputs = pipeline(instruction, max_new_tokens=512, do_sample=True, temperature=0.9)
   print(outputs[0]["generated_text"])
   

4. Hardware Requirements: Due to its size, it is recommended to use cloud GPUs such as those available on AWS, Google Cloud, or Azure for optimal performance.

License

The WIROAI-TURKISH-LLM-9B model is released under Google's Gemma license. Users must review and accept the license terms before utilizing the model.