Add new CrossEncoder model
Browse files- README.md +66 -66
- config.json +46 -46
- onnx/model.onnx +3 -0
- special_tokens_map.json +37 -37
- tokenizer_config.json +58 -57
README.md
CHANGED
|
@@ -1,67 +1,67 @@
|
|
| 1 |
-
---
|
| 2 |
-
license: apache-2.0
|
| 3 |
-
pipeline_tag: text-ranking
|
| 4 |
-
language:
|
| 5 |
-
- en
|
| 6 |
-
library_name: sentence-transformers
|
| 7 |
-
base_model:
|
| 8 |
-
- google/electra-base-discriminator
|
| 9 |
-
tags:
|
| 10 |
-
- transformers
|
| 11 |
-
---
|
| 12 |
-
|
| 13 |
-
## Cross-Encoder for Text Ranking
|
| 14 |
-
|
| 15 |
-
This model is a port of the [webis/monoelectra-base](https://huggingface.co/webis/monoelectra-base) model from [lightning-ir](https://github.com/webis-de/lightning-ir) to [Sentence Transformers](https://sbert.net/) and [Transformers](https://huggingface.co/docs/transformers).
|
| 16 |
-
|
| 17 |
-
The original model was introduced in the paper [A Systematic Investigation of Distilling Large Language Models into Cross-Encoders for Passage Re-ranking](https://arxiv.org/abs/2405.07920). See https://github.com/webis-de/rank-distillm for code used to train the original model.
|
| 18 |
-
|
| 19 |
-
The model can be used as a reranker in a 2-stage "retrieve-rerank" pipeline, where it reorders passages returned by a retriever model (e.g. an embedding model or BM25) given some query. See [SBERT.net Retrieve & Re-rank](https://www.sbert.net/examples/applications/retrieve_rerank/README.html) for more details.
|
| 20 |
-
|
| 21 |
-
## Usage with Sentence Transformers
|
| 22 |
-
|
| 23 |
-
The usage is easy when you have [SentenceTransformers](https://www.sbert.net/) installed.
|
| 24 |
-
|
| 25 |
-
```bash
|
| 26 |
-
pip install sentence-transformers
|
| 27 |
-
```
|
| 28 |
-
|
| 29 |
-
Then you can use the pre-trained model like this:
|
| 30 |
-
|
| 31 |
-
```python
|
| 32 |
-
from sentence_transformers import CrossEncoder
|
| 33 |
-
|
| 34 |
-
model = CrossEncoder("cross-encoder/monoelectra-base", trust_remote_code=True)
|
| 35 |
-
scores = model.predict([
|
| 36 |
-
("How many people live in Berlin?", "Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."),
|
| 37 |
-
("How many people live in Berlin?", "Berlin is well known for its museums."),
|
| 38 |
-
])
|
| 39 |
-
print(scores)
|
| 40 |
-
# [ 8.122868 -4.292924]
|
| 41 |
-
```
|
| 42 |
-
|
| 43 |
-
## Usage with Transformers
|
| 44 |
-
|
| 45 |
-
```python
|
| 46 |
-
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
| 47 |
-
import torch
|
| 48 |
-
|
| 49 |
-
model = AutoModelForSequenceClassification.from_pretrained("cross-encoder/monoelectra-base", trust_remote_code=True)
|
| 50 |
-
tokenizer = AutoTokenizer.from_pretrained("cross-encoder/monoelectra-base")
|
| 51 |
-
|
| 52 |
-
features = tokenizer(
|
| 53 |
-
[
|
| 54 |
-
("How many people live in Berlin?", "Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."),
|
| 55 |
-
("How many people live in Berlin?", "Berlin is well known for its museums."),
|
| 56 |
-
],
|
| 57 |
-
padding=True,
|
| 58 |
-
truncation=True,
|
| 59 |
-
return_tensors="pt",
|
| 60 |
-
)
|
| 61 |
-
|
| 62 |
-
model.eval()
|
| 63 |
-
with torch.no_grad():
|
| 64 |
-
scores = model(**features).logits.view(-1)
|
| 65 |
-
print(scores)
|
| 66 |
-
# tensor([ 8.1229, -4.2929])
|
| 67 |
```
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
pipeline_tag: text-ranking
|
| 4 |
+
language:
|
| 5 |
+
- en
|
| 6 |
+
library_name: sentence-transformers
|
| 7 |
+
base_model:
|
| 8 |
+
- google/electra-base-discriminator
|
| 9 |
+
tags:
|
| 10 |
+
- transformers
|
| 11 |
+
---
|
| 12 |
+
|
| 13 |
+
## Cross-Encoder for Text Ranking
|
| 14 |
+
|
| 15 |
+
This model is a port of the [webis/monoelectra-base](https://huggingface.co/webis/monoelectra-base) model from [lightning-ir](https://github.com/webis-de/lightning-ir) to [Sentence Transformers](https://sbert.net/) and [Transformers](https://huggingface.co/docs/transformers).
|
| 16 |
+
|
| 17 |
+
The original model was introduced in the paper [A Systematic Investigation of Distilling Large Language Models into Cross-Encoders for Passage Re-ranking](https://arxiv.org/abs/2405.07920). See https://github.com/webis-de/rank-distillm for code used to train the original model.
|
| 18 |
+
|
| 19 |
+
The model can be used as a reranker in a 2-stage "retrieve-rerank" pipeline, where it reorders passages returned by a retriever model (e.g. an embedding model or BM25) given some query. See [SBERT.net Retrieve & Re-rank](https://www.sbert.net/examples/applications/retrieve_rerank/README.html) for more details.
|
| 20 |
+
|
| 21 |
+
## Usage with Sentence Transformers
|
| 22 |
+
|
| 23 |
+
The usage is easy when you have [SentenceTransformers](https://www.sbert.net/) installed.
|
| 24 |
+
|
| 25 |
+
```bash
|
| 26 |
+
pip install sentence-transformers
|
| 27 |
+
```
|
| 28 |
+
|
| 29 |
+
Then you can use the pre-trained model like this:
|
| 30 |
+
|
| 31 |
+
```python
|
| 32 |
+
from sentence_transformers import CrossEncoder
|
| 33 |
+
|
| 34 |
+
model = CrossEncoder("cross-encoder/monoelectra-base", trust_remote_code=True)
|
| 35 |
+
scores = model.predict([
|
| 36 |
+
("How many people live in Berlin?", "Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."),
|
| 37 |
+
("How many people live in Berlin?", "Berlin is well known for its museums."),
|
| 38 |
+
])
|
| 39 |
+
print(scores)
|
| 40 |
+
# [ 8.122868 -4.292924]
|
| 41 |
+
```
|
| 42 |
+
|
| 43 |
+
## Usage with Transformers
|
| 44 |
+
|
| 45 |
+
```python
|
| 46 |
+
from transformers import AutoTokenizer, AutoModelForSequenceClassification
|
| 47 |
+
import torch
|
| 48 |
+
|
| 49 |
+
model = AutoModelForSequenceClassification.from_pretrained("cross-encoder/monoelectra-base", trust_remote_code=True)
|
| 50 |
+
tokenizer = AutoTokenizer.from_pretrained("cross-encoder/monoelectra-base")
|
| 51 |
+
|
| 52 |
+
features = tokenizer(
|
| 53 |
+
[
|
| 54 |
+
("How many people live in Berlin?", "Berlin had a population of 3,520,031 registered inhabitants in an area of 891.82 square kilometers."),
|
| 55 |
+
("How many people live in Berlin?", "Berlin is well known for its museums."),
|
| 56 |
+
],
|
| 57 |
+
padding=True,
|
| 58 |
+
truncation=True,
|
| 59 |
+
return_tensors="pt",
|
| 60 |
+
)
|
| 61 |
+
|
| 62 |
+
model.eval()
|
| 63 |
+
with torch.no_grad():
|
| 64 |
+
scores = model(**features).logits.view(-1)
|
| 65 |
+
print(scores)
|
| 66 |
+
# tensor([ 8.1229, -4.2929])
|
| 67 |
```
|
config.json
CHANGED
|
@@ -1,46 +1,46 @@
|
|
| 1 |
-
{
|
| 2 |
-
"architectures": [
|
| 3 |
-
"WebisCrossEncoderForSequenceClassification"
|
| 4 |
-
],
|
| 5 |
-
"attention_probs_dropout_prob": 0.1,
|
| 6 |
-
"auto_map": {
|
| 7 |
-
"AutoModelForSequenceClassification": "modeling.WebisCrossEncoderForSequenceClassification"
|
| 8 |
-
},
|
| 9 |
-
"backbone_model_type": "electra",
|
| 10 |
-
"classifier_dropout": null,
|
| 11 |
-
"doc_length": 256,
|
| 12 |
-
"embedding_size": 768,
|
| 13 |
-
"hidden_act": "gelu",
|
| 14 |
-
"hidden_dropout_prob": 0.1,
|
| 15 |
-
"hidden_size": 768,
|
| 16 |
-
"id2label": {
|
| 17 |
-
"0": "LABEL_0"
|
| 18 |
-
},
|
| 19 |
-
"initializer_range": 0.02,
|
| 20 |
-
"intermediate_size": 3072,
|
| 21 |
-
"label2id": {
|
| 22 |
-
"LABEL_0": 0
|
| 23 |
-
},
|
| 24 |
-
"layer_norm_eps": 1e-12,
|
| 25 |
-
"max_position_embeddings": 512,
|
| 26 |
-
"model_type": "electra",
|
| 27 |
-
"num_attention_heads": 12,
|
| 28 |
-
"num_hidden_layers": 12,
|
| 29 |
-
"pad_token_id": 0,
|
| 30 |
-
"pooling_strategy": "first",
|
| 31 |
-
"position_embedding_type": "absolute",
|
| 32 |
-
"query_length": 32,
|
| 33 |
-
"sentence_transformers": {
|
| 34 |
-
"activation_fn": "torch.nn.modules.linear.Identity",
|
| 35 |
-
"version": "4.0.
|
| 36 |
-
},
|
| 37 |
-
"summary_activation": "gelu",
|
| 38 |
-
"summary_last_dropout": 0.1,
|
| 39 |
-
"summary_type": "first",
|
| 40 |
-
"summary_use_proj": true,
|
| 41 |
-
"torch_dtype": "float32",
|
| 42 |
-
"transformers_version": "4.
|
| 43 |
-
"type_vocab_size": 2,
|
| 44 |
-
"use_cache": true,
|
| 45 |
-
"vocab_size": 30522
|
| 46 |
-
}
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"WebisCrossEncoderForSequenceClassification"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"auto_map": {
|
| 7 |
+
"AutoModelForSequenceClassification": "cross-encoder/monoelectra-base--modeling.WebisCrossEncoderForSequenceClassification"
|
| 8 |
+
},
|
| 9 |
+
"backbone_model_type": "electra",
|
| 10 |
+
"classifier_dropout": null,
|
| 11 |
+
"doc_length": 256,
|
| 12 |
+
"embedding_size": 768,
|
| 13 |
+
"hidden_act": "gelu",
|
| 14 |
+
"hidden_dropout_prob": 0.1,
|
| 15 |
+
"hidden_size": 768,
|
| 16 |
+
"id2label": {
|
| 17 |
+
"0": "LABEL_0"
|
| 18 |
+
},
|
| 19 |
+
"initializer_range": 0.02,
|
| 20 |
+
"intermediate_size": 3072,
|
| 21 |
+
"label2id": {
|
| 22 |
+
"LABEL_0": 0
|
| 23 |
+
},
|
| 24 |
+
"layer_norm_eps": 1e-12,
|
| 25 |
+
"max_position_embeddings": 512,
|
| 26 |
+
"model_type": "electra",
|
| 27 |
+
"num_attention_heads": 12,
|
| 28 |
+
"num_hidden_layers": 12,
|
| 29 |
+
"pad_token_id": 0,
|
| 30 |
+
"pooling_strategy": "first",
|
| 31 |
+
"position_embedding_type": "absolute",
|
| 32 |
+
"query_length": 32,
|
| 33 |
+
"sentence_transformers": {
|
| 34 |
+
"activation_fn": "torch.nn.modules.linear.Identity",
|
| 35 |
+
"version": "4.1.0.dev0"
|
| 36 |
+
},
|
| 37 |
+
"summary_activation": "gelu",
|
| 38 |
+
"summary_last_dropout": 0.1,
|
| 39 |
+
"summary_type": "first",
|
| 40 |
+
"summary_use_proj": true,
|
| 41 |
+
"torch_dtype": "float32",
|
| 42 |
+
"transformers_version": "4.52.0.dev0",
|
| 43 |
+
"type_vocab_size": 2,
|
| 44 |
+
"use_cache": true,
|
| 45 |
+
"vocab_size": 30522
|
| 46 |
+
}
|
onnx/model.onnx
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aef62e9a40e6a7e84a1b5e3c5e20c2aea3d8dc6a67bb9ed6581f0eb91823b6a0
|
| 3 |
+
size 438212375
|
special_tokens_map.json
CHANGED
|
@@ -1,37 +1,37 @@
|
|
| 1 |
-
{
|
| 2 |
-
"cls_token": {
|
| 3 |
-
"content": "[CLS]",
|
| 4 |
-
"lstrip": false,
|
| 5 |
-
"normalized": false,
|
| 6 |
-
"rstrip": false,
|
| 7 |
-
"single_word": false
|
| 8 |
-
},
|
| 9 |
-
"mask_token": {
|
| 10 |
-
"content": "[MASK]",
|
| 11 |
-
"lstrip": false,
|
| 12 |
-
"normalized": false,
|
| 13 |
-
"rstrip": false,
|
| 14 |
-
"single_word": false
|
| 15 |
-
},
|
| 16 |
-
"pad_token": {
|
| 17 |
-
"content": "[PAD]",
|
| 18 |
-
"lstrip": false,
|
| 19 |
-
"normalized": false,
|
| 20 |
-
"rstrip": false,
|
| 21 |
-
"single_word": false
|
| 22 |
-
},
|
| 23 |
-
"sep_token": {
|
| 24 |
-
"content": "[SEP]",
|
| 25 |
-
"lstrip": false,
|
| 26 |
-
"normalized": false,
|
| 27 |
-
"rstrip": false,
|
| 28 |
-
"single_word": false
|
| 29 |
-
},
|
| 30 |
-
"unk_token": {
|
| 31 |
-
"content": "[UNK]",
|
| 32 |
-
"lstrip": false,
|
| 33 |
-
"normalized": false,
|
| 34 |
-
"rstrip": false,
|
| 35 |
-
"single_word": false
|
| 36 |
-
}
|
| 37 |
-
}
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"cls_token": {
|
| 3 |
+
"content": "[CLS]",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"mask_token": {
|
| 10 |
+
"content": "[MASK]",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"pad_token": {
|
| 17 |
+
"content": "[PAD]",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"sep_token": {
|
| 24 |
+
"content": "[SEP]",
|
| 25 |
+
"lstrip": false,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"unk_token": {
|
| 31 |
+
"content": "[UNK]",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
}
|
| 37 |
+
}
|
tokenizer_config.json
CHANGED
|
@@ -1,57 +1,58 @@
|
|
| 1 |
-
{
|
| 2 |
-
"added_tokens_decoder": {
|
| 3 |
-
"0": {
|
| 4 |
-
"content": "[PAD]",
|
| 5 |
-
"lstrip": false,
|
| 6 |
-
"normalized": false,
|
| 7 |
-
"rstrip": false,
|
| 8 |
-
"single_word": false,
|
| 9 |
-
"special": true
|
| 10 |
-
},
|
| 11 |
-
"100": {
|
| 12 |
-
"content": "[UNK]",
|
| 13 |
-
"lstrip": false,
|
| 14 |
-
"normalized": false,
|
| 15 |
-
"rstrip": false,
|
| 16 |
-
"single_word": false,
|
| 17 |
-
"special": true
|
| 18 |
-
},
|
| 19 |
-
"101": {
|
| 20 |
-
"content": "[CLS]",
|
| 21 |
-
"lstrip": false,
|
| 22 |
-
"normalized": false,
|
| 23 |
-
"rstrip": false,
|
| 24 |
-
"single_word": false,
|
| 25 |
-
"special": true
|
| 26 |
-
},
|
| 27 |
-
"102": {
|
| 28 |
-
"content": "[SEP]",
|
| 29 |
-
"lstrip": false,
|
| 30 |
-
"normalized": false,
|
| 31 |
-
"rstrip": false,
|
| 32 |
-
"single_word": false,
|
| 33 |
-
"special": true
|
| 34 |
-
},
|
| 35 |
-
"103": {
|
| 36 |
-
"content": "[MASK]",
|
| 37 |
-
"lstrip": false,
|
| 38 |
-
"normalized": false,
|
| 39 |
-
"rstrip": false,
|
| 40 |
-
"single_word": false,
|
| 41 |
-
"special": true
|
| 42 |
-
}
|
| 43 |
-
},
|
| 44 |
-
"clean_up_tokenization_spaces": true,
|
| 45 |
-
"cls_token": "[CLS]",
|
| 46 |
-
"do_lower_case": true,
|
| 47 |
-
"doc_length": 256,
|
| 48 |
-
"
|
| 49 |
-
"
|
| 50 |
-
"
|
| 51 |
-
"
|
| 52 |
-
"
|
| 53 |
-
"
|
| 54 |
-
"
|
| 55 |
-
"
|
| 56 |
-
"
|
| 57 |
-
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "[PAD]",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"100": {
|
| 12 |
+
"content": "[UNK]",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"101": {
|
| 20 |
+
"content": "[CLS]",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"102": {
|
| 28 |
+
"content": "[SEP]",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"103": {
|
| 36 |
+
"content": "[MASK]",
|
| 37 |
+
"lstrip": false,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"clean_up_tokenization_spaces": true,
|
| 45 |
+
"cls_token": "[CLS]",
|
| 46 |
+
"do_lower_case": true,
|
| 47 |
+
"doc_length": 256,
|
| 48 |
+
"extra_special_tokens": {},
|
| 49 |
+
"mask_token": "[MASK]",
|
| 50 |
+
"model_max_length": 512,
|
| 51 |
+
"pad_token": "[PAD]",
|
| 52 |
+
"query_length": 32,
|
| 53 |
+
"sep_token": "[SEP]",
|
| 54 |
+
"strip_accents": null,
|
| 55 |
+
"tokenize_chinese_chars": true,
|
| 56 |
+
"tokenizer_class": "ElectraTokenizer",
|
| 57 |
+
"unk_token": "[UNK]"
|
| 58 |
+
}
|