florian-hoenicke commited on
Commit
00c615e
1 Parent(s): 03d2ac0

feat: push custom model

Browse files
1_Pooling/config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "word_embedding_dimension": 768,
3
+ "pooling_mode_cls_token": true,
4
+ "pooling_mode_mean_tokens": false,
5
+ "pooling_mode_max_tokens": false,
6
+ "pooling_mode_mean_sqrt_len_tokens": false,
7
+ "pooling_mode_weightedmean_tokens": false,
8
+ "pooling_mode_lasttoken": false,
9
+ "include_prompt": true
10
+ }
README.md ADDED
@@ -0,0 +1,41 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ datasets:
4
+ - fine-tuned/arguana-c-256-24-gpt-4o-2024-05-13-51550
5
+ - allenai/c4
6
+ language:
7
+ - en
8
+ pipeline_tag: feature-extraction
9
+ tags:
10
+ - sentence-transformers
11
+ - feature-extraction
12
+ - sentence-similarity
13
+ - mteb
14
+ - Argumentation
15
+ - Corpus
16
+ - Research
17
+ - Quality
18
+ - Dataset
19
+ ---
20
+ This model is a fine-tuned version of [**jinaai/jina-embeddings-v2-base-en**](https://huggingface.co/jinaai/jina-embeddings-v2-base-en) designed for the following use case:
21
+
22
+ academic research data retrieval
23
+
24
+ ## How to Use
25
+ This model can be easily integrated into your NLP pipeline for tasks such as text classification, sentiment analysis, entity recognition, and more. Here's a simple example to get you started:
26
+
27
+ ```python
28
+ from sentence_transformers import SentenceTransformer
29
+ from sentence_transformers.util import cos_sim
30
+
31
+ model = SentenceTransformer(
32
+ 'fine-tuned/arguana-c-256-24-gpt-4o-2024-05-13-51550',
33
+ trust_remote_code=True
34
+ )
35
+
36
+ embeddings = model.encode([
37
+ 'first text to embed',
38
+ 'second text to embed'
39
+ ])
40
+ print(cos_sim(embeddings[0], embeddings[1]))
41
+ ```
config.json ADDED
@@ -0,0 +1,36 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "_name_or_path": "mytmp/finetuned_model",
3
+ "architectures": [
4
+ "JinaBertModel"
5
+ ],
6
+ "attention_probs_dropout_prob": 0.0,
7
+ "attn_implementation": null,
8
+ "auto_map": {
9
+ "AutoConfig": "configuration_bert.JinaBertConfig",
10
+ "AutoModel": "modeling_bert.JinaBertModel",
11
+ "AutoModelForMaskedLM": "jinaai/jina-bert-implementation--modeling_bert.JinaBertForMaskedLM",
12
+ "AutoModelForSequenceClassification": "jinaai/jina-bert-implementation--modeling_bert.JinaBertForSequenceClassification"
13
+ },
14
+ "classifier_dropout": null,
15
+ "emb_pooler": "mean",
16
+ "feed_forward_type": "geglu",
17
+ "gradient_checkpointing": false,
18
+ "hidden_act": "gelu",
19
+ "hidden_dropout_prob": 0.1,
20
+ "hidden_size": 768,
21
+ "initializer_range": 0.02,
22
+ "intermediate_size": 3072,
23
+ "layer_norm_eps": 1e-12,
24
+ "max_position_embeddings": 8192,
25
+ "model_max_length": 8192,
26
+ "model_type": "bert",
27
+ "num_attention_heads": 12,
28
+ "num_hidden_layers": 12,
29
+ "pad_token_id": 0,
30
+ "position_embedding_type": "alibi",
31
+ "torch_dtype": "float32",
32
+ "transformers_version": "4.40.2",
33
+ "type_vocab_size": 2,
34
+ "use_cache": true,
35
+ "vocab_size": 30528
36
+ }
config_sentence_transformers.json ADDED
@@ -0,0 +1,9 @@
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "__version__": {
3
+ "sentence_transformers": "2.7.0",
4
+ "transformers": "4.40.2",
5
+ "pytorch": "1.11.0+cpu"
6
+ },
7
+ "prompts": {},
8
+ "default_prompt_name": null
9
+ }
configuration_bert.py ADDED
@@ -0,0 +1,168 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ # coding=utf-8
2
+ # Copyright 2018 The Google AI Language Team Authors and The HuggingFace Inc. team.
3
+ # Copyright (c) 2018, NVIDIA CORPORATION. All rights reserved.
4
+ # Copyright (c) 2023 Jina AI GmbH. All rights reserved.
5
+ #
6
+ # Licensed under the Apache License, Version 2.0 (the "License");
7
+ # you may not use this file except in compliance with the License.
8
+ # You may obtain a copy of the License at
9
+ #
10
+ # http://www.apache.org/licenses/LICENSE-2.0
11
+ #
12
+ # Unless required by applicable law or agreed to in writing, software
13
+ # distributed under the License is distributed on an "AS IS" BASIS,
14
+ # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
15
+ # See the License for the specific language governing permissions and
16
+ # limitations under the License.
17
+ """ BERT model configuration"""
18
+ from collections import OrderedDict
19
+ from typing import Mapping
20
+
21
+ from transformers.configuration_utils import PretrainedConfig
22
+ from transformers.onnx import OnnxConfig
23
+ from transformers.utils import logging
24
+
25
+
26
+ logger = logging.get_logger(__name__)
27
+
28
+
29
+ class JinaBertConfig(PretrainedConfig):
30
+ r"""
31
+ This is the configuration class to store the configuration of a [`JinaBertModel`]. It is used to
32
+ instantiate a BERT model according to the specified arguments, defining the model architecture. Instantiating a
33
+ configuration with the defaults will yield a similar configuration to that of the BERT
34
+ [bert-base-uncased](https://huggingface.co/bert-base-uncased) architecture.
35
+
36
+ Configuration objects inherit from [`PretrainedConfig`] and can be used to control the model outputs. Read the
37
+ documentation from [`PretrainedConfig`] for more information.
38
+
39
+
40
+ Args:
41
+ vocab_size (`int`, *optional*, defaults to 30522):
42
+ Vocabulary size of the BERT model. Defines the number of different tokens that can be represented by the
43
+ `inputs_ids` passed when calling [`BertModel`] or [`TFBertModel`].
44
+ hidden_size (`int`, *optional*, defaults to 768):
45
+ Dimensionality of the encoder layers and the pooler layer.
46
+ num_hidden_layers (`int`, *optional*, defaults to 12):
47
+ Number of hidden layers in the Transformer encoder.
48
+ num_attention_heads (`int`, *optional*, defaults to 12):
49
+ Number of attention heads for each attention layer in the Transformer encoder.
50
+ intermediate_size (`int`, *optional*, defaults to 3072):
51
+ Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder.
52
+ hidden_act (`str` or `Callable`, *optional*, defaults to `"gelu"`):
53
+ The non-linear activation function (function or string) in the encoder and pooler. If string, `"gelu"`,
54
+ `"relu"`, `"silu"` and `"gelu_new"` are supported.
55
+ hidden_dropout_prob (`float`, *optional*, defaults to 0.1):
56
+ The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
57
+ attention_probs_dropout_prob (`float`, *optional*, defaults to 0.1):
58
+ The dropout ratio for the attention probabilities.
59
+ max_position_embeddings (`int`, *optional*, defaults to 512):
60
+ The maximum sequence length that this model might ever be used with. Typically set this to something large
61
+ just in case (e.g., 512 or 1024 or 2048).
62
+ type_vocab_size (`int`, *optional*, defaults to 2):
63
+ The vocabulary size of the `token_type_ids` passed when calling [`BertModel`] or [`TFBertModel`].
64
+ initializer_range (`float`, *optional*, defaults to 0.02):
65
+ The standard deviation of the truncated_normal_initializer for initializing all weight matrices.
66
+ layer_norm_eps (`float`, *optional*, defaults to 1e-12):
67
+ The epsilon used by the layer normalization layers.
68
+ position_embedding_type (`str`, *optional*, defaults to `"absolute"`):
69
+ Type of position embedding. Choose one of `"absolute"`, `"relative_key"`, `"relative_key_query"`. For
70
+ positional embeddings use `"absolute"`. For more information on `"relative_key"`, please refer to
71
+ [Self-Attention with Relative Position Representations (Shaw et al.)](https://arxiv.org/abs/1803.02155).
72
+ For more information on `"relative_key_query"`, please refer to *Method 4* in [Improve Transformer Models
73
+ with Better Relative Position Embeddings (Huang et al.)](https://arxiv.org/abs/2009.13658).
74
+ is_decoder (`bool`, *optional*, defaults to `False`):
75
+ Whether the model is used as a decoder or not. If `False`, the model is used as an encoder.
76
+ use_cache (`bool`, *optional*, defaults to `True`):
77
+ Whether or not the model should return the last key/values attentions (not used by all models). Only
78
+ relevant if `config.is_decoder=True`.
79
+ classifier_dropout (`float`, *optional*):
80
+ The dropout ratio for the classification head.
81
+ feed_forward_type (`str`, *optional*, defaults to `"original"`):
82
+ The type of feed forward layer to use in the bert layers.
83
+ Can be one of GLU variants, e.g. `"reglu"`, `"geglu"`
84
+ emb_pooler (`str`, *optional*, defaults to `None`):
85
+ The function to use for pooling the last layer embeddings to get the sentence embeddings.
86
+ Should be one of `None`, `"mean"`.
87
+ attn_implementation (`str`, *optional*, defaults to `"torch"`):
88
+ The implementation of the self-attention layer. Can be one of:
89
+ - `None` for the original implementation,
90
+ - `torch` for the PyTorch SDPA implementation,
91
+
92
+ Examples:
93
+
94
+ ```python
95
+ >>> from transformers import JinaBertConfig, JinaBertModel
96
+
97
+ >>> # Initializing a JinaBert configuration
98
+ >>> configuration = JinaBertConfig()
99
+
100
+ >>> # Initializing a model (with random weights) from the configuration
101
+ >>> model = JinaBertModel(configuration)
102
+
103
+ >>> # Accessing the model configuration
104
+ >>> configuration = model.config
105
+
106
+ >>> # Encode text inputs
107
+ >>> embeddings = model.encode(text_inputs)
108
+ ```"""
109
+ model_type = "bert"
110
+
111
+ def __init__(
112
+ self,
113
+ vocab_size=30522,
114
+ hidden_size=768,
115
+ num_hidden_layers=12,
116
+ num_attention_heads=12,
117
+ intermediate_size=3072,
118
+ hidden_act="gelu",
119
+ hidden_dropout_prob=0.1,
120
+ attention_probs_dropout_prob=0.1,
121
+ max_position_embeddings=512,
122
+ type_vocab_size=2,
123
+ initializer_range=0.02,
124
+ layer_norm_eps=1e-12,
125
+ pad_token_id=0,
126
+ position_embedding_type="absolute",
127
+ use_cache=True,
128
+ classifier_dropout=None,
129
+ feed_forward_type="original",
130
+ emb_pooler=None,
131
+ attn_implementation='torch',
132
+ **kwargs,
133
+ ):
134
+ super().__init__(pad_token_id=pad_token_id, **kwargs)
135
+
136
+ self.vocab_size = vocab_size
137
+ self.hidden_size = hidden_size
138
+ self.num_hidden_layers = num_hidden_layers
139
+ self.num_attention_heads = num_attention_heads
140
+ self.hidden_act = hidden_act
141
+ self.intermediate_size = intermediate_size
142
+ self.hidden_dropout_prob = hidden_dropout_prob
143
+ self.attention_probs_dropout_prob = attention_probs_dropout_prob
144
+ self.max_position_embeddings = max_position_embeddings
145
+ self.type_vocab_size = type_vocab_size
146
+ self.initializer_range = initializer_range
147
+ self.layer_norm_eps = layer_norm_eps
148
+ self.position_embedding_type = position_embedding_type
149
+ self.use_cache = use_cache
150
+ self.classifier_dropout = classifier_dropout
151
+ self.feed_forward_type = feed_forward_type
152
+ self.emb_pooler = emb_pooler
153
+ self.attn_implementation = attn_implementation
154
+
155
+ class JinaBertOnnxConfig(OnnxConfig):
156
+ @property
157
+ def inputs(self) -> Mapping[str, Mapping[int, str]]:
158
+ if self.task == "multiple-choice":
159
+ dynamic_axis = {0: "batch", 1: "choice", 2: "sequence"}
160
+ else:
161
+ dynamic_axis = {0: "batch", 1: "sequence"}
162
+ return OrderedDict(
163
+ [
164
+ ("input_ids", dynamic_axis),
165
+ ("attention_mask", dynamic_axis),
166
+ ("token_type_ids", dynamic_axis),
167
+ ]
168
+ )
model.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:5b8d59bd0cd67ac1286851d3313d13d5b683cf1777752cacba47dfc5dd55ff70
3
+ size 549493968
modeling_bert.py ADDED
The diff for this file is too large to render. See raw diff
 
modules.json ADDED
@@ -0,0 +1,20 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ [
2
+ {
3
+ "idx": 0,
4
+ "name": "0",
5
+ "path": "",
6
+ "type": "sentence_transformers.models.Transformer"
7
+ },
8
+ {
9
+ "idx": 1,
10
+ "name": "1",
11
+ "path": "1_Pooling",
12
+ "type": "sentence_transformers.models.Pooling"
13
+ },
14
+ {
15
+ "idx": 2,
16
+ "name": "2",
17
+ "path": "2_Normalize",
18
+ "type": "sentence_transformers.models.Normalize"
19
+ }
20
+ ]
sentence_bert_config.json ADDED
@@ -0,0 +1,4 @@
 
 
 
 
 
1
+ {
2
+ "max_seq_length": 8192,
3
+ "do_lower_case": false
4
+ }
special_tokens_map.json ADDED
@@ -0,0 +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
+ }
tokenizer.json ADDED
The diff for this file is too large to render. See raw diff
 
tokenizer_config.json ADDED
@@ -0,0 +1,57 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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_basic_tokenize": true,
47
+ "do_lower_case": true,
48
+ "mask_token": "[MASK]",
49
+ "model_max_length": 2147483648,
50
+ "never_split": null,
51
+ "pad_token": "[PAD]",
52
+ "sep_token": "[SEP]",
53
+ "strip_accents": null,
54
+ "tokenize_chinese_chars": true,
55
+ "tokenizer_class": "BertTokenizer",
56
+ "unk_token": "[UNK]"
57
+ }
training_args.bin ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:3b29932be9ad0a40b84fab055579a241bccbd372a4d681581f3ff61f803f1632
3
+ size 4719
vocab.txt ADDED
The diff for this file is too large to render. See raw diff