File size: 17,163 Bytes
005cdaa
 
 
 
 
 
bcb052a
 
005cdaa
 
b687b75
b4b186c
b687b75
b4b186c
b687b75
 
 
 
 
b4b186c
b687b75
 
 
b4b186c
 
b687b75
b4b186c
b687b75
 
 
 
 
c2997b7
 
b4b186c
 
b687b75
 
b4b186c
b687b75
b4b186c
b687b75
b4b186c
b687b75
b4b186c
b687b75
 
b4b186c
b687b75
 
 
b4b186c
b687b75
b4b186c
b687b75
b4b186c
b687b75
 
 
b4b186c
b687b75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b602448
 
b687b75
 
519af80
b687b75
 
519af80
b687b75
 
 
 
 
 
b602448
 
b687b75
 
519af80
b687b75
 
519af80
b687b75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b602448
b687b75
 
 
 
 
 
 
 
 
 
 
 
 
 
c0d5b82
 
 
 
 
 
 
 
 
 
 
 
 
 
b687b75
 
 
 
 
 
 
 
c2997b7
b687b75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c2997b7
b687b75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c2997b7
b687b75
 
 
 
b4b186c
 
 
b687b75
b4b186c
f788030
b4b186c
 
b687b75
b4b186c
b687b75
b4b186c
f788030
b4b186c
b687b75
b4b186c
f788030
b4b186c
b687b75
 
b4b186c
f788030
b4b186c
b687b75
 
b4b186c
f788030
b4b186c
 
 
b687b75
b4b186c
bcb052a
b4b186c
bcb052a
b687b75
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bcb052a
 
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
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
---
license: apache-2.0
pipeline_tag: text-classification
tags:
- transformers
- sentence-transformers
language:
- multilingual
---

# Reranker

**More details please refer to our Github: [FlagEmbedding](https://github.com/FlagOpen/FlagEmbedding/tree/master).**

- [Model List](#model-list)
- [Usage](#usage)
- [Fine-tuning](#fine-tune)
- [Evaluation](#evaluation)
- [Citation](#citation)

Different from embedding model, reranker uses question and document as input and directly output similarity instead of embedding. 
You can get a relevance score by inputting query and passage to the reranker. 
And the score can be mapped to a float value in [0,1] by sigmoid function.


## Model List

| Model                                                                     | Base model                                                           | Language | layerwise |                           feature                            |
|:--------------------------------------------------------------------------|:--------:|:-----------------------------------------------------------------------------------------------------------------------------------:|:----------------------------------------------------------------------------------------------:|:----------------------------------------------------------------------------------------------:|
| [BAAI/bge-reranker-base](https://huggingface.co/BAAI/bge-reranker-base) | [xlm-roberta-base](https://huggingface.co/xlm-roberta-base) | Chinese and English |     -     | Lightweight reranker model, easy to deploy, with fast inference. |
| [BAAI/bge-reranker-large](https://huggingface.co/BAAI/bge-reranker-large) | [xlm-roberta-large](https://huggingface.co/FacebookAI/xlm-roberta-large) | Chinese and English |     -     | Lightweight reranker model, easy to deploy, with fast inference. |
| [BAAI/bge-reranker-v2-m3](https://huggingface.co/BAAI/bge-reranker-v2-m3) | [bge-m3](https://huggingface.co/BAAI/bge-m3) |    Multilingual     |     -     | Lightweight reranker model, possesses strong multilingual capabilities, easy to deploy, with fast inference. |
| [BAAI/bge-reranker-v2-gemma](https://huggingface.co/BAAI/bge-reranker-v2-gemma) |      [gemma-2b](https://huggingface.co/google/gemma-2b)      |    Multilingual     |     -     | Suitable for multilingual contexts, performs well in both English proficiency and multilingual capabilities. |
| [BAAI/bge-reranker-v2-minicpm-layerwise](https://huggingface.co/BAAI/bge-reranker-v2-minicpm-layerwise) | [MiniCPM-2B-dpo-bf16](https://huggingface.co/openbmb/MiniCPM-2B-dpo-bf16) |    Multilingual     |   8-40    | Suitable for multilingual contexts, performs well in both English and Chinese proficiency, allows freedom to select layers for output, facilitating accelerated inference. |


You can select the model according your senario and resource. 
- For **multilingual**, utilize [BAAI/bge-reranker-v2-m3](https://huggingface.co/BAAI/bge-reranker-v2-m3) and [BAAI/bge-reranker-v2-gemma](https://huggingface.co/BAAI/bge-reranker-v2-gemma)

- For **Chinese or English**, utilize [BAAI/bge-reranker-v2-m3](https://huggingface.co/BAAI/bge-reranker-v2-m3) and [BAAI/bge-reranker-v2-minicpm-layerwise](https://huggingface.co/BAAI/bge-reranker-v2-minicpm-layerwise). 

- For **efficiency**, utilize [BAAI/bge-reranker-v2-m3](https://huggingface.co/BAAI/bge-reranker-v2-m3) and the low layer of [BAAI/bge-reranker-v2-minicpm-layerwise](https://huggingface.co/BAAI/bge-reranker-v2-minicpm-layerwise). 

- For better performance, recommand [BAAI/bge-reranker-v2-minicpm-layerwise](https://huggingface.co/BAAI/bge-reranker-v2-minicpm-layerwise) and [BAAI/bge-reranker-v2-gemma](https://huggingface.co/BAAI/bge-reranker-v2-gemma)

## Usage 
### Using FlagEmbedding

```
pip install -U FlagEmbedding
```

#### For normal reranker (bge-reranker-base / bge-reranker-large / bge-reranker-v2-m3 )

Get relevance scores (higher scores indicate more relevance):

```python
from FlagEmbedding import FlagReranker
reranker = FlagReranker('BAAI/bge-reranker-v2-m3', use_fp16=True) # Setting use_fp16 to True speeds up computation with a slight performance degradation

score = reranker.compute_score(['query', 'passage'])
print(score) # -5.65234375

# You can map the scores into 0-1 by set "normalize=True", which will apply sigmoid function to the score
score = reranker.compute_score(['query', 'passage'], normalize=True)
print(score) # 0.003497010252573502

scores = reranker.compute_score([['what is panda?', 'hi'], ['what is panda?', 'The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China.']])
print(scores) # [-8.1875, 5.26171875]

# You can map the scores into 0-1 by set "normalize=True", which will apply sigmoid function to the score
scores = reranker.compute_score([['what is panda?', 'hi'], ['what is panda?', 'The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China.']], normalize=True)
print(scores) # [0.00027803096387751553, 0.9948403768236574]
```

#### For LLM-based reranker

```python
from FlagEmbedding import FlagLLMReranker
reranker = FlagLLMReranker('BAAI/bge-reranker-v2-gemma', use_fp16=True) # Setting use_fp16 to True speeds up computation with a slight performance degradation
# reranker = FlagLLMReranker('BAAI/bge-reranker-v2-gemma', use_bf16=True) # You can also set use_bf16=True to speed up computation with a slight performance degradation

score = reranker.compute_score(['query', 'passage'])
print(score)

scores = reranker.compute_score([['what is panda?', 'hi'], ['what is panda?', 'The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China.']])
print(scores)
```

#### For LLM-based layerwise reranker

```python
from FlagEmbedding import LayerWiseFlagLLMReranker
reranker = LayerWiseFlagLLMReranker('BAAI/bge-reranker-v2-minicpm-layerwise', use_fp16=True) # Setting use_fp16 to True speeds up computation with a slight performance degradation
# reranker = LayerWiseFlagLLMReranker('BAAI/bge-reranker-v2-minicpm-layerwise', use_bf16=True) # You can also set use_bf16=True to speed up computation with a slight performance degradation

score = reranker.compute_score(['query', 'passage'], cutoff_layers=[28]) # Adjusting 'cutoff_layers' to pick which layers are used for computing the score.
print(score)

scores = reranker.compute_score([['what is panda?', 'hi'], ['what is panda?', 'The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China.']], cutoff_layers=[28])
print(scores)
```

### Using Huggingface transformers

#### For normal reranker (bge-reranker-base / bge-reranker-large / bge-reranker-v2-m3 )

Get relevance scores (higher scores indicate more relevance):

```python
import torch
from transformers import AutoModelForSequenceClassification, AutoTokenizer

tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-reranker-v2-m3')
model = AutoModelForSequenceClassification.from_pretrained('BAAI/bge-reranker-v2-m3')
model.eval()

pairs = [['what is panda?', 'hi'], ['what is panda?', 'The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China.']]
with torch.no_grad():
    inputs = tokenizer(pairs, padding=True, truncation=True, return_tensors='pt', max_length=512)
    scores = model(**inputs, return_dict=True).logits.view(-1, ).float()
    print(scores)
```

#### For LLM-based reranker

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

def get_inputs(pairs, tokenizer, prompt=None, max_length=1024):
    if prompt is None:
        prompt = "Given a query A and a passage B, determine whether the passage contains an answer to the query by providing a prediction of either 'Yes' or 'No'."
    sep = "\n"
    prompt_inputs = tokenizer(prompt,
                              return_tensors=None,
                              add_special_tokens=False)['input_ids']
    sep_inputs = tokenizer(sep,
                           return_tensors=None,
                           add_special_tokens=False)['input_ids']
    inputs = []
    for query, passage in pairs:
        query_inputs = tokenizer(f'A: {query}',
                                 return_tensors=None,
                                 add_special_tokens=False,
                                 max_length=max_length * 3 // 4,
                                 truncation=True)
        passage_inputs = tokenizer(f'B: {passage}',
                                   return_tensors=None,
                                   add_special_tokens=False,
                                   max_length=max_length,
                                   truncation=True)
        item = tokenizer.prepare_for_model(
            [tokenizer.bos_token_id] + query_inputs['input_ids'],
            sep_inputs + passage_inputs['input_ids'],
            truncation='only_second',
            max_length=max_length,
            padding=False,
            return_attention_mask=False,
            return_token_type_ids=False,
            add_special_tokens=False
        )
        item['input_ids'] = item['input_ids'] + sep_inputs + prompt_inputs
        item['attention_mask'] = [1] * len(item['input_ids'])
        inputs.append(item)
    return tokenizer.pad(
            inputs,
            padding=True,
            max_length=max_length + len(sep_inputs) + len(prompt_inputs),
            pad_to_multiple_of=8,
            return_tensors='pt',
    )

tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-reranker-v2-gemma')
model = AutoModelForCausalLM.from_pretrained('BAAI/bge-reranker-v2-gemma')
yes_loc = tokenizer('Yes', add_special_tokens=False)['input_ids'][0]
model.eval()

pairs = [['what is panda?', 'hi'], ['what is panda?', 'The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China.']]
with torch.no_grad():
    inputs = get_inputs(pairs, tokenizer)
    scores = model(**inputs, return_dict=True).logits[:, -1, yes_loc].view(-1, ).float()
    print(scores)
```

#### For LLM-based layerwise reranker

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

def get_inputs(pairs, tokenizer, prompt=None, max_length=1024):
    if prompt is None:
        prompt = "Given a query A and a passage B, determine whether the passage contains an answer to the query by providing a prediction of either 'Yes' or 'No'."
    sep = "\n"
    prompt_inputs = tokenizer(prompt,
                              return_tensors=None,
                              add_special_tokens=False)['input_ids']
    sep_inputs = tokenizer(sep,
                           return_tensors=None,
                           add_special_tokens=False)['input_ids']
    inputs = []
    for query, passage in pairs:
        query_inputs = tokenizer(f'A: {query}',
                                 return_tensors=None,
                                 add_special_tokens=False,
                                 max_length=max_length * 3 // 4,
                                 truncation=True)
        passage_inputs = tokenizer(f'B: {passage}',
                                   return_tensors=None,
                                   add_special_tokens=False,
                                   max_length=max_length,
                                   truncation=True)
        item = tokenizer.prepare_for_model(
            [tokenizer.bos_token_id] + query_inputs['input_ids'],
            sep_inputs + passage_inputs['input_ids'],
            truncation='only_second',
            max_length=max_length,
            padding=False,
            return_attention_mask=False,
            return_token_type_ids=False,
            add_special_tokens=False
        )
        item['input_ids'] = item['input_ids'] + sep_inputs + prompt_inputs
        item['attention_mask'] = [1] * len(item['input_ids'])
        inputs.append(item)
    return tokenizer.pad(
            inputs,
            padding=True,
            max_length=max_length + len(sep_inputs) + len(prompt_inputs),
            pad_to_multiple_of=8,
            return_tensors='pt',
    )

tokenizer = AutoTokenizer.from_pretrained('BAAI/bge-reranker-v2-minicpm-layerwise', trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained('BAAI/bge-reranker-v2-minicpm-layerwise', trust_remote_code=True, torch_dtype=torch.bfloat16)
model = model.to('cuda')
model.eval()

pairs = [['what is panda?', 'hi'], ['what is panda?', 'The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China.']]
with torch.no_grad():
    inputs = get_inputs(pairs, tokenizer).to(model.device)
    all_scores = model(**inputs, return_dict=True, cutoff_layers=[28])
    all_scores = [scores[:, -1].view(-1, ).float() for scores in all_scores[0]]
    print(all_scores)
```

## Fine-tune

### Data Format

Train data should be a json file, where each line is a dict like this:

```
{"query": str, "pos": List[str], "neg":List[str], "prompt": str}
```

`query` is the query, and `pos` is a list of positive texts, `neg` is a list of negative texts, `prompt` indicates the relationship between query and texts. If you have no negative texts for a query, you can random sample some from the entire corpus as the negatives.

See [toy_finetune_data.jsonl](https://github.com/FlagOpen/FlagEmbedding/tree/master/FlagEmbedding/llm_reranker/toy_finetune_data.jsonl) for a toy data file.

### Train

You can fine-tune the reranker with the following code:

**For llm-based reranker**

```shell
torchrun --nproc_per_node {number of gpus} \
-m FlagEmbedding.llm_reranker.finetune_for_instruction.run \
--output_dir {path to save model} \
--model_name_or_path google/gemma-2b \
--train_data ./toy_finetune_data.jsonl \
--learning_rate 2e-4 \
--num_train_epochs 1 \
--per_device_train_batch_size 1 \
--gradient_accumulation_steps 16 \
--dataloader_drop_last True \
--query_max_len 512 \
--passage_max_len 512 \
--train_group_size 16 \
--logging_steps 1 \
--save_steps 2000 \
--save_total_limit 50 \
--ddp_find_unused_parameters False \
--gradient_checkpointing \
--deepspeed stage1.json \
--warmup_ratio 0.1 \
--bf16 \
--use_lora True \
--lora_rank 32 \
--lora_alpha 64 \
--use_flash_attn True \
--target_modules q_proj k_proj v_proj o_proj
```

**For llm-based layerwise reranker**

```shell
torchrun --nproc_per_node {number of gpus} \
-m FlagEmbedding.llm_reranker.finetune_for_layerwise.run \
--output_dir {path to save model} \
--model_name_or_path openbmb/MiniCPM-2B-dpo-bf16 \
--train_data ./toy_finetune_data.jsonl \
--learning_rate 2e-4 \
--num_train_epochs 1 \
--per_device_train_batch_size 1 \
--gradient_accumulation_steps 16 \
--dataloader_drop_last True \
--query_max_len 512 \
--passage_max_len 512 \
--train_group_size 16 \
--logging_steps 1 \
--save_steps 2000 \
--save_total_limit 50 \
--ddp_find_unused_parameters False \
--gradient_checkpointing \
--deepspeed stage1.json \
--warmup_ratio 0.1 \
--bf16 \
--use_lora True \
--lora_rank 32 \
--lora_alpha 64 \
--use_flash_attn True \
--target_modules q_proj k_proj v_proj o_proj \
--start_layer 8 \
--head_multi True \
--head_type simple
```

Our rerankers are initialized from [google/gemma-2b](https://huggingface.co/google/gemma-2b) (for llm-based reranker) and [openbmb/MiniCPM-2B-dpo-bf16](https://huggingface.co/openbmb/MiniCPM-2B-dpo-bf16) (for llm-based layerwise reranker), and we train it on a mixture of multilingual datasets:

- [bge-m3-data](https://huggingface.co/datasets/Shitao/bge-m3-data)
- [quora train data](https://huggingface.co/datasets/quora)
- [fever train data](https://fever.ai/dataset/fever.html)

## Evaluation

- llama-index.

![image-20240317193909373](./assets/llama-index.png)


- BEIR.   

rereank the top 100 results from bge-en-v1.5 large.

![image-20240317174633333](./assets/BEIR-bge-en-v1.5.png)

rereank the top 100 results from e5 mistral 7b instruct.

![image-20240317172949713](./assets/BEIR-e5-mistral.png)

- CMTEB-retrieval.   
It rereank the top 100 results from bge-zh-v1.5 large.

![image-20240317173026235](./assets/CMTEB-retrieval-bge-zh-v1.5.png)

- miracl (multi-language).   
It rereank the top 100 results from bge-m3.

![image-20240317173117639](./assets/miracl-bge-m3.png)



## Citation

If you find this repository useful, please consider giving a star and citation

```bibtex
@misc{li2023making,
      title={Making Large Language Models A Better Foundation For Dense Retrieval}, 
      author={Chaofan Li and Zheng Liu and Shitao Xiao and Yingxia Shao},
      year={2023},
      eprint={2312.15503},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
@misc{chen2024bge,
      title={BGE M3-Embedding: Multi-Lingual, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation}, 
      author={Jianlv Chen and Shitao Xiao and Peitian Zhang and Kun Luo and Defu Lian and Zheng Liu},
      year={2024},
      eprint={2402.03216},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}
```