File size: 1,702 Bytes
ec25a6e
1240f58
 
 
ec25a6e
 
70b712a
1240f58
ec25a6e
1240f58
ec25a6e
1240f58
ec25a6e
 
1240f58
 
 
ec25a6e
1240f58
ec25a6e
1240f58
ec25a6e
1240f58
 
ec25a6e
1240f58
0d8b104
1240f58
 
 
 
ec25a6e
1240f58
 
 
 
 
 
 
 
 
 
 
ec25a6e
1240f58
 
 
 
 
 
 
ec25a6e
1240f58
ec25a6e
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
---
license: apache-2.0
language:
- tr
---

<img src="https://huggingface.co./Orbina/Orbita-v0.1/resolve/main/orbita.png"
alt="Orbita LLM" width="500"/>

# 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

```python

from transformers import AutoModelForCausalLM, AutoTokenizer
device = "cuda" # the device to load the model onto

model = AutoModelForCausalLM.from_pretrained(
    "Orbina/Orbita-v0.1",
    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]