|
--- |
|
license: apache-2.0 |
|
--- |
|
|
|
<img src="https://huggingface.co./datasets/hkust-nlp/deita-images/resolve/main/logo-final.png" alt="Deita banner" width="800" style="margin-left:'auto' margin-right:'auto' display:'block'"/> |
|
|
|
# Model Card for Deita Quality Scorer |
|
|
|
[GitHub](https://github.com/hkust-nlp/deita) | [Paper](https://arxiv.org/abs/2312.15685) |
|
|
|
Deita is an open-sourced project designed to facilitate Automatic Data Selection for instruction tuning in Large Language Models (LLMs). |
|
|
|
Deita Quality Scorer is a tool for automatically annotating the Instruction Quality of SFT data. |
|
|
|
## Model description |
|
|
|
- **Model type:** Model fine tuned to automatically annotate the Instruction-Response Pair Quality |
|
- **Language(s) (NLP):** Primarily English |
|
- **Finetuned from model:** Llama-1-13b-hf |
|
|
|
### Model Sources |
|
|
|
- **Repository:** https://github.com/hkust-nlp/deita |
|
- **Model Family:** Other models and the dataset are found in the [Deita collection](https://huggingface.co./collections/hkust-nlp/deita-6569c198c174808d94cf5bd4). |
|
|
|
|
|
## Performance |
|
|
|
|
|
|
|
| Model | Align | Data Size | MT-Bench | AlpacaEval(%) | OpenLLM (Avg.) | |
|
|------------------------------------------------|-----------|------------|----------|---------------|----------------| |
|
| **Proprietary Models** | | | | | | |
|
| GPT-4-Turbo | ? | -- | 9.32 | 97.70 | -- | |
|
| GPT-4 | SFT + PPO | -- | 8.99 | 95.03 | -- | |
|
| Claude-2 | SFT + PPO | -- | 8.06 | 91.36 | -- | |
|
| GPT-3.5-turbo | SFT + PPO | -- | 7.94 | 89.37 | -- | |
|
| **Open-sourced Models based on LLaMA-1-13B** | | | | | | |
|
| LIMA | SFT | 1K SFT | 4.29 | 41.98 | 59.82 | |
|
| WizardLM-13B | SFT | 70K SFT | 6.35 | 75.31 | 58.96 | |
|
| Vicuna-13B-v1.3 | SFT | 125K SFT | 6.39 | 82.11 | 60.01 | |
|
| Random | SFT | 10K SFT | 6.03 | 71.52 | 60.14 | |
|
| DEITA-LLaMA1-13B-v1.0-sft | SFT | 10K SFT | 6.60 | 78.01 | 64.27 | |
|
| **Open-sourced Models based on LLaMA-2-13B** | | | | | | |
|
| Tulu-2-13B | SFT | 326K SFT | 6.70 | 78.90 | -- | |
|
| Tulu-2-13B+DPO | SFT + DPO | 326K SFT + 60K DPO | 7.00 | 89.50 | -- | |
|
| LLaMA2-13B-Chat | SFT + PPO | -- | 6.65 | 81.09 | -- | |
|
| WizardLM-13B-v1.2 | SFT | >70K SFT | 7.09 | 89.17 | -- | |
|
| Vicuna-13B-v1.5 | SFT | 125K SFT | 6.57 | 78.80 | 61.63 | |
|
| Random | SFT | 10K SFT | 5.78 | 65.19 | 61.32 | |
|
| DEITA-LLaMA2-13B-v1.0-sft | SFT | 10K SFT | 6.79 | 81.09 | 62.71 | |
|
| **Open-sourced Models based on Mistral-7B** | | | | | | |
|
| Mistral-7B-Instruct-v0.1 | -- | -- | 6.84 | 69.65 | 60.45 | |
|
| Zephyr-7B-sft | SFT | 200K SFT | 5.32 | 75.12 | 60.93 | |
|
| $\text{Zephyr-7B-}\beta$ | SFT + DPO | 200K SFT + 60K DPO | 7.34 | 90.60 | 66.36 | |
|
| OpenChat-3.5 | C-RLFT | >> 70K C-RLFT | 7.81 | 88.51 | -- | |
|
| Starling-7B | C-RLFT + APA | >>70K C-RLFT + 183K APA | 8.09 | 91.99 | -- | |
|
| Random | SFT | 10K SFT | 5.89 | 56.90 | 61.72 | |
|
| DEITA-7B-v1.0-sft (6K) | SFT | 6K SFT | 7.22 | 80.78 | 64.94 | |
|
| DEITA-7B-v1.0-sft (10K) | SFT | 10K SFT | 7.32 | 81.67 | 64.00 | |
|
| DEITA-7B-v1.0 | SFT + DPO | 6K SFT + 10K DPO | 7.55 | 90.06 | 69.86 | |
|
|
|
|
|
|
|
|
|
|
|
## Usage |
|
|
|
Please use the following format to score the quality of Instruction-Response Pair |
|
|
|
```python |
|
from transformers import AutoTokenizer, AutoModelForCausalLM |
|
import numpy as np |
|
from scipy.special import softmax |
|
model_name = "hkust-nlp/deita-quality-scorer" |
|
tokenizer = AutoTokenizer.from_pretrained(model_name) |
|
model = AutoModelForCausalLM.from_pretrained(model_name) |
|
|
|
|
|
def infer_Quality(model, tokenizer, input_text, resp_text): |
|
quality_template = ("You are a helpful assistant. Please identify the quality score of the Response corresponding to the Question. \n #Question#:\n{instruction}\n#Response#:\n{output} \n##Quality: ") |
|
user_input = quality_template.format(instruction=input_text, output=resp_text) |
|
input_ids = tokenizer.encode(user_input, return_tensors="pt") |
|
max_length = 512 |
|
outputs = model.generate(input_ids, max_length=512, num_return_sequences=1, return_dict_in_generate=True, output_scores=True) |
|
logprobs_list = outputs.scores[0][0] |
|
score_logits = [] |
|
id2score = { |
|
29896: "1", |
|
29906: "2", |
|
29941: "3", |
|
29946: "4", |
|
29945: "5", |
|
29953: "6" |
|
} |
|
score_template = np.array([1,2,3,4,5,6]) |
|
for k in id2score: |
|
score_logits.append(logprobs_list[k]) |
|
score_logits = np.array(score_logits) |
|
score_npy = softmax(score_logits, axis=0) |
|
score_npy = score_npy * score_template |
|
|
|
score_npy = np.sum(score_npy, axis=0) |
|
return score_npy |
|
|
|
input_text = "word to describe UI with helpful tooltips" # Example Input |
|
output_text = "User-friendly or intuitive UI" # Example Output |
|
quality_score = infer_quality(model, tokenizer, input_text) |
|
|
|
print(quality_score) |
|
|
|
|
|
``` |
|
|
|
## Citation |
|
If you find the content of this project helpful, please cite our paper as follows: |
|
|
|
``` |
|
@misc{liu2023what, |
|
title={What Makes Good Data for Alignment? A Comprehensive Study of Automatic Data Selection in Instruction Tuning}, |
|
author={Wei Liu and Weihao Zeng and Keqing He and Yong Jiang and Junxian He}, |
|
year={2023}, |
|
eprint={2312.15685}, |
|
archivePrefix={arXiv}, |
|
primaryClass={cs.CL} |
|
} |
|
``` |
|
|
|
|
