Kona2-12B-Instruct is a 12-billion parameter instruction-tuned language model for Georgian and English. Built on
Kona2-12B-Base through supervised fine-tuning (SFT), it excels at chat, question answering, and function calling.
1from transformers import AutoModelForCausalLM, AutoTokenizer
2
3model = AutoModelForCausalLM.from_pretrained(
4 "tbilisi-ai-lab/kona2-12B-Instruct",
5 torch_dtype="auto",
6 device_map="auto"
7)
8tokenizer = AutoTokenizer.from_pretrained("tbilisi-ai-lab/kona2-12B-Instruct")
9
10messages = [
11 {"role": "system", "content": "You are a helpful assistant."},
12 {"role": "user", "content": "რა არის საქართველოს დედაქალაქი?"}
13]
14
15inputs = tokenizer.apply_chat_template(
16 messages,
17 return_tensors="pt",
18 add_generation_prompt=True
19).to(model.device)
20
21outputs = model.generate(inputs, max_new_tokens=256, temperature=0.7)
22print(tokenizer.decode(outputs[0], skip_special_tokens=True))
1tools = [
2 {
3 "type": "function",
4 "function": {
5 "name": "get_weather",
6 "description": "Get current weather for a location",
7 "parameters": {
8 "type": "object",
9 "properties": {
10 "location": {"type": "string", "description": "City name"}
11 },
12 "required": ["location"]
13 }
14 }
15 }
16]
17
18messages = [
19 {"role": "system", "content": "You are a helpful assistant with access to tools."},
20 {"role": "user", "content": "What's the weather in Tbilisi?"}
21]
22
23inputs = tokenizer.apply_chat_template(
24 messages,
25 tools=tools,
26 return_tensors="pt",
27 add_generation_prompt=True
28).to(model.device)
29
30outputs = model.generate(inputs, max_new_tokens=256)
31# Output will include <tool_call>{"name": "get_weather", "arguments": {"location": "Tbilisi"}}</tool_call>
1@misc{tbilisi2025kona2instruct,
2 title = {Kona2-12B-Instruct: A Georgian Instruction-Tuned Language Model},
3 author = {Tbilisi AI Lab Team},
4 year = {2025},
5 publisher = {Hugging Face},
6 howpublished = {\url{https://huggingface.co/tbilisi-ai-lab/kona2-12B-Instruct}}
7}
This model is released under the
Apache 2.0 License.
