metadata
license: apache-2.0
language:
- tr
Orbita-v0.1
This model is an extended version of a Qwen-based Large Language Model (LLM) for Turkish. It was trained on a cleaned Turkish dataset carefully annotated to carry out turkish instructions in an accurate and organized manner. This model was fully finetuned extensively on 8 H100 GPU's for 2 days using a carefully annotated Turkish dataset.
Model Details
- Base Model: Qwen 14B based LLM
- Training Dataset: Annotated Turkish Dataset
- Training Method: Full Finetuning
Usage Examples
from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda" # the device to load the model onto
model = AutoModelForCausalLM.from_pretrained(
"Qwen/Qwen1.5-14B-Chat",
torch_dtype="auto",
device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen1.5-14B-Chat")
prompt = "türkiyenin inflasyonu nasıl çözebiliriz?"
messages = [
{"role": "system", "content": "Sen Orbina ai tarafından üretelen bir yapay zekasındır, soruları uygun bir şekilde cevap veriyorsun"},
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(device)
generated_ids = model.generate(
model_inputs.input_ids,
max_new_tokens=512
)
generated_ids = [
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]