Rustamshry/Qwen2.5-3B-Self-Instruct-Turkish

Model Card for Model ID

Model Description

This model is a fine-tuned version of Qwen2.5-3B-Instruct, optimized for Turkish instruction-following tasks. Leveraging the CausalLM/GPT-4-Self-Instruct-Turkish dataset, the model has been trained to understand and respond to a wide range of Turkish prompts, enhancing its capabilities in tasks such as question answering

• Language(s) (NLP): Turkish
• License: MIT
• Finetuned from model: unsloth/Qwen2.5-3B-Instruct

Uses

Direct Use

This model is intended for applications requiring Turkish language understanding and generation, particularly in instruction-following scenarios.

How to Get Started with the Model

Use the code below to get started with the model.

from huggingface_hub import login
from transformers import AutoTokenizer, AutoModelForCausalLM
from peft import PeftModel

login(token="")

tokenizer = AutoTokenizer.from_pretrained("unsloth/Qwen2.5-3B-Instruct",)
base_model = AutoModelForCausalLM.from_pretrained(
    "unsloth/Qwen2.5-3B-Instruct",
    device_map="auto", token=""
)

model = PeftModel.from_pretrained(base_model,"Rustamshry/Qwen2.5-3B-Self-Instruct-Turkish")


question = "Türkiye'deki sağlık hizmetleri ve hastaneler hakkında genel bir özet oluşturun."

prompt = (
    f"### Soru:\n{question}\n\n"
    f"### Cevap:\n"
)

input_ids = tokenizer(prompt, return_tensors="pt").to(model.device)

outputs = model.generate(
    **input_ids,
    max_new_tokens=2048,
    #temperature=0.6,
    #top_p=0.95,
    #do_sample=True,
    #eos_token_id=tokenizer.eos_token_id
)

print(tokenizer.decode(outputs[0]),skip_special_tokens=True)

Training Details

Training Data

• Dataset: CausalLM/GPT-4-Self-Instruct-Turkish

• Description: A collection of Turkish instruction-response pairs generated using the Self-Instruct methodology, where GPT-4 was employed to create synthetic instruction data. This approach aims to improve the model's ability to follow diverse and complex instructions in Turkish.

Framework versions

• PEFT 0.15.2