AhiskaAI-134m-Instruct-v0.2

AhiskaAI-134m-IT-v0.2

AhiskaAI-134m-IT-v0.2 is the instruction-tuned version of our 134M parameter Small Language Model. This model has been fine-tuned on 16,000+ high-quality, curated Turkish instruction-response pairs to function as a helpful and conversational AI assistant.

Base Model: AhiskaAI-134m-Base-v0.2

Model Details

• Architecture: Llama-based architecture.
• Fine-tuning: Supervised Fine-Tuning (SFT) on 16k+ instruction pairs.
• Format: ChatML.
• Parameters: 134M.
• Hardware: Trained on NVIDIA RTX 4050 Laptop GPU.

Training Logs

Usage (ChatML Format)

This model is optimized for chat interactions. Please use the following ChatML structure for best results:

Recommended System Prompt

To get the best performance, use the following system prompt: "Sen kibar, sorulan soruları tam cümlelerle yanıtlayan Türkçe bir asistansın."

Example Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model = AutoModelForCausalLM.from_pretrained("AhiskaAI/AhiskaAI-134m-IT-v0.2")
tokenizer = AutoTokenizer.from_pretrained("AhiskaAI/AhiskaAI-134m-IT-v0.2")

SYSTEM_PROMPT = "Sen kibar, sorulan soruları tam cümlelerle yanıtlayan Türkçe bir asistansın."
user_query = "Ahıska Türkleri hakkında bilgi verir misin?"

prompt = (
    f"<|im_start|>system\n{SYSTEM_PROMPT}<|im_end|>\n"
    f"<|im_start|>user\n{user_query}<|im_end|>\n"
    f"<|im_start|>assistant\n"
)

inputs = tokenizer(prompt, return_tensors="pt")
outputs = model.generate(**inputs, max_new_tokens=200)
print(tokenizer.decode(outputs[0], skip_special_tokens=True))