File size: 2,850 Bytes
d61eccf c5197a1 |
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 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 |
---
language:
- ko
- en
pipeline_tag: text-generation
tags:
- orca
- llama-2
---
<p align="center" width="100%">
<img src="https://i.imgur.com/snFDU0P.png" alt="KoreanLM icon" style="width: 500px; display: block; margin: auto; border-radius: 10%;">
</p>
# quantumaikr/llama-2-70b-fb16-korean
## Model Description
`quantumaikr/llama-2-70b-fb16-korean` is a Llama2 70B model finetuned the Korean Dataset
## Usage
Start chatting with `quantumaikr/llama-2-70b-fb16-korean` using the following code snippet:
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer, pipeline
tokenizer = AutoTokenizer.from_pretrained("quantumaikr/llama-2-70b-fb16-korean")
model = AutoModelForCausalLM.from_pretrained("quantumaikr/llama-2-70b-fb16-korean", torch_dtype=torch.float16, device_map="auto")
system_prompt = "### System:\n๊ทํ๋ ์ง์๋ฅผ ๋งค์ฐ ์ ๋ฐ๋ฅด๋ AI์ธ QuantumLM์
๋๋ค. ์ต๋ํ ๋ง์ด ๋์์ฃผ์ธ์. ์์ ์ ์ ์ํ๊ณ ๋ถ๋ฒ์ ์ธ ํ๋์ ํ์ง ๋ง์ธ์.\n\n"
message = "์ธ๊ณต์ง๋ฅ์ด๋ ๋ฌด์์ธ๊ฐ์?"
prompt = f"{system_prompt}### User: {message}\n\n### Assistant:\n"
inputs = tokenizer(prompt, return_tensors="pt").to("cuda")
output = model.generate(**inputs, do_sample=True, temperature=0.9, top_p=0.75, max_new_tokens=4096)
print(tokenizer.decode(output[0], skip_special_tokens=True))
```
QuantumLM should be used with this prompt format:
```
### System:
This is a system prompt, please behave and help the user.
### User:
Your prompt here
### Assistant
The output of QuantumLM
```
## Use and Limitations
### Intended Use
These models are intended for research only, in adherence with the [CC BY-NC-4.0](https://creativecommons.org/licenses/by-nc/4.0/) license.
### Limitations and bias
Although the aforementioned dataset helps to steer the base language models into "safer" distributions of text, not all biases and toxicity can be mitigated through fine-tuning. We ask that users be mindful of such potential issues that can arise in generated responses. Do not treat model outputs as substitutes for human judgment or as sources of truth. Please use it responsibly.
Contact us : hi@quantumai.kr
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_quantumaikr__llama-2-70b-fb16-korean)
| Metric | Value |
|-----------------------|---------------------------|
| Avg. | 56.97 |
| ARC (25-shot) | 67.15 |
| HellaSwag (10-shot) | 86.78 |
| MMLU (5-shot) | 69.29 |
| TruthfulQA (0-shot) | 56.5 |
| Winogrande (5-shot) | 82.64 |
| GSM8K (5-shot) | 29.04 |
| DROP (3-shot) | 7.42 |
|