Llama-2-7b-chat-tr / README.md
malhajar's picture
Update README.md
dc63145
|
raw
history blame
2.07 kB
---
library_name: peft
base_model: meta-llama/Llama-2-7b-hf
language:
- en
- tr
tags:
- llama-2
- turkish
- dolly
datasets:
- atasoglu/databricks-dolly-15k-tr
---
# Model Card for Model ID
<!-- Provide a quick summary of what the model is/does. -->
malhajar/Llama-2-7b-chat-dolly-tr is a finetuned version of Llama-2-7b-hf using SFT Training.
This model can answer information in turkish language as it is finetuned on a turkish dataset specifically [`databricks-dolly-15k-tr`]( https://huggingface.co/datasets/atasoglu/databricks-dolly-15k-tr)
![llama](./llama.png)
### Model Description
- **Developed by:** [`Mohamad Alhajar`](https://www.linkedin.com/in/muhammet-alhajar/)
- **Language(s) (NLP):** Turkish
- **Finetuned from model:** [`meta-llama/Llama-2-7b-hf`](https://huggingface.co/meta-llama/Llama-2-7b-hf)
### Prompt Template
```
<s>[INST] <prompt> [/INST]
```
## How to Get Started with the Model
Use the code sample provided in the original post to interact with the model.
```python
from transformers import AutoTokenizer,AutoModelForCausalLM
model_id = "malhajar/Llama-2-7b-chat-dolly-tr"
model = AutoModelForCausalLM.from_pretrained(model_name_or_path,
device_map="auto",
torch_dtype=torch.float16,
revision="main")
tokenizer = AutoTokenizer.from_pretrained(model_id)
question: "Türkiyenin en büyük şehir nedir?"
# For generating a response
prompt = '''
<s>[INST] {question} [/INST]
'''
input_ids = tokenizer(prompt, return_tensors="pt").input_ids
output = model.generate(inputs=input_ids,max_new_tokens=512,pad_token_id=tokenizer.eos_token_id,top_k=50, do_sample=True,repetition_penalty=1.3
top_p=0.95)
response = tokenizer.decode(output[0])
print(response)
```
## Example Generation
```
<s>[INST] Türkiyenin en büyük şehir nedir? [/INST]
İstanbul, dünyanın en kalabalık ikinci ve Turuncu kütle'de yer almaktadır. Pek çok insandaki birçok ünlüsün bulundusuyla biliniyor.
```