File size: 4,184 Bytes
77dee57
 
 
9e6c5d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
07f62b0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
9e6c5d4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
---


# RankingGPT-bloom-3b

RankingGPT is a text ranker based on large language models with significant in-domain and out-domain effectiveness.
We provide RankingGPT in different sizes and types, including bloom-560m, bloom-1b1, bloom-3b, bloom-7b, llama2-7b, baichuan2-7b and qwen-7b.

More details please refer to our [paper](https://arxiv.org/abs/2311.16720) and [github](https://github.com/Alibaba-NLP/RankingGPT).


## Usage

Code example
```python
import torch
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained('zyznull/RankingGPT-bloom-3b')
model = AutoModelForCausalLM.from_pretrained('zyznull/RankingGPT-bloom-3b').eval()

query='when should a baby walk'
document='Most babies start to walk around 13 months, but your baby may start walking as early as 9 or 10 months or as late as 15 or 16 months.'

context=f'Document: {document} Query:'
example=context+query

context_enc = tokenizer.encode(context, add_special_tokens=False)
continuation_enc = tokenizer.encode(query, add_special_tokens=False)
model_input = torch.tensor(context_enc+continuation_enc[:-1])
continuation_len = len(continuation_enc)
input_len, = model_input.shape


with torch.no_grad():
    logprobs = torch.nn.functional.log_softmax(model(model_input.unsqueeze(dim=0))[0], dim=-1)[0]

logprobs = logprobs[input_len-continuation_len:]
logprobs = torch.gather(logprobs, 1, torch.tensor(continuation_enc).unsqueeze(-1)).squeeze(-1)
score = torch.sum(logprobs)/logprobs.shape[0]

print(f"Document: {document[:20] + '...'} Score: {score}")
```

### Result
|         | DL19 | DL20 | BEIR | url |
|---------|------|------|------|-----------------|
| MonoBERT-340M | 72.3 | 70.3 | 50.5 |     [huggingface](https://huggingface.co/veneres/monobert-msmarco)          |
| MonoT5-220M  | 71.5 | 69.7 | 49.3 |     [huggingface](https://huggingface.co/castorini/monot5-base-msmarco)          |
| MonoT5-770M  | 73.2 | 71.2 | 53.1 |    [huggingface](https://huggingface.co/castorini/monot5-large-msmarco)          |
| MonoT5-3B  | 72.8 | 74.5 | 54.6 |     [huggingface](https://huggingface.co/castorini/monot5-3b-msmarco)          |
| RankT5-770M  | -    | -    | 53.7 |     [huggingface](https://huggingface.co/bergum/rank-T5-flan)           |
| RankLLaMA| 74.6 | 76.6 | 52.5 |  [huggingface](https://huggingface.co/castorini/rankllama-v1-7b-lora-passage) |
| RankingGPT-bloom-560m| 75.3 | 73.2 | 53.7 | [huggingface](https://huggingface.co/zyznull/RankingGPT-bloom-560m) [modelscope](https://modelscope.cn/models/damo/RankingGPT-bloom-560m)       |
| RankingGPT-bloom-1b1| 75.6 | 73.2 | 54.5 | [huggingface](https://huggingface.co/zyznull/RankingGPT-bloom-1b1)  [modelscope](https://modelscope.cn/models/damo/RankingGPT-bloom-1b1)        |
| RankingGPT-bloom-3b| 76.8 | 73.6 | 56.2 | [huggingface](https://huggingface.co/zyznull/RankingGPT-bloom-3b)  [modelscope](https://modelscope.cn/models/damo/RankingGPT-bloom-3b)        |
| RankingGPT-bloom-7b| 77.3 | 74.6 | 56.6 | [huggingface](https://huggingface.co/zyznull/RankingGPT-bloom-7b)  [modelscope](https://modelscope.cn/models/damo/RankingGPT-bloom-7b)        |
| RankingGPT-llama2-7b| 76.2 | 76.3 | 57.8 | [huggingface](https://huggingface.co/zyznull/RankingGPT-llama2-7b)  [modelscope](https://modelscope.cn/models/damo/RankingGPT-llama2-7b)        |
| RankingGPT-baichuan2-7b| 75.9 | 74.3 | 57.5 |  [huggingface](https://huggingface.co/zyznull/RankingGPT-baichuan2-7b) [modelscope](https://modelscope.cn/models/damo/RankingGPT-baichuan2-7b)        |
| RankingGPT-qwen-7b| 75.8 | 74.3 | 58.3 | [huggingface](https://huggingface.co/zyznull/RankingGPT-qwen-7b)  [modelscope](https://modelscope.cn/models/damo/RankingGPT-qwen-7b)        |

### Citation

If you find our paper or models helpful, please consider citing them as follows:

```
@misc{zhang2023rankinggpt,
      title={RankingGPT: Empowering Large Language Models in Text Ranking with Progressive Enhancement}, 
      author={Longhui Zhang and Yanzhao Zhang and Dingkun Long and Pengjun Xie and Meishan Zhang and Min Zhang},
      year={2023},
      eprint={2311.16720},
      archivePrefix={arXiv},
      primaryClass={cs.IR}
}
```