Latest training run, just 4 epochs, optimizations all pulled except for FP16, save and eval at epochs to avoid over-fitting
Browse filesThis view is limited to 50 files because it contains too many changes.
See raw diff
- 1_Pooling/config.json +10 -0
- checkpoint-1386/1_Pooling/config.json +10 -0
- checkpoint-1386/README.md +460 -0
- checkpoint-1386/config.json +25 -0
- checkpoint-1386/config_sentence_transformers.json +10 -0
- checkpoint-1386/model.safetensors +3 -0
- checkpoint-1386/modules.json +14 -0
- checkpoint-1386/optimizer.pt +3 -0
- checkpoint-1386/rng_state.pth +3 -0
- checkpoint-1386/scaler.pt +3 -0
- checkpoint-1386/scheduler.pt +3 -0
- checkpoint-1386/sentence_bert_config.json +4 -0
- checkpoint-1386/special_tokens_map.json +51 -0
- checkpoint-1386/tokenizer.json +3 -0
- checkpoint-1386/tokenizer_config.json +65 -0
- checkpoint-1386/trainer_state.json +89 -0
- checkpoint-1386/training_args.bin +3 -0
- checkpoint-1386/unigram.json +3 -0
- checkpoint-2079/1_Pooling/config.json +10 -0
- checkpoint-2079/README.md +463 -0
- checkpoint-2079/config.json +25 -0
- checkpoint-2079/config_sentence_transformers.json +10 -0
- checkpoint-2079/model.safetensors +3 -0
- checkpoint-2079/modules.json +14 -0
- checkpoint-2079/optimizer.pt +3 -0
- checkpoint-2079/rng_state.pth +3 -0
- checkpoint-2079/scaler.pt +3 -0
- checkpoint-2079/scheduler.pt +3 -0
- checkpoint-2079/sentence_bert_config.json +4 -0
- checkpoint-2079/special_tokens_map.json +51 -0
- checkpoint-2079/tokenizer.json +3 -0
- checkpoint-2079/tokenizer_config.json +65 -0
- checkpoint-2079/trainer_state.json +119 -0
- checkpoint-2079/training_args.bin +3 -0
- checkpoint-2079/unigram.json +3 -0
- checkpoint-2772/1_Pooling/config.json +10 -0
- checkpoint-2772/README.md +465 -0
- checkpoint-2772/config.json +25 -0
- checkpoint-2772/config_sentence_transformers.json +10 -0
- checkpoint-2772/model.safetensors +3 -0
- checkpoint-2772/modules.json +14 -0
- checkpoint-2772/optimizer.pt +3 -0
- checkpoint-2772/rng_state.pth +3 -0
- checkpoint-2772/scaler.pt +3 -0
- checkpoint-2772/scheduler.pt +3 -0
- checkpoint-2772/sentence_bert_config.json +4 -0
- checkpoint-2772/special_tokens_map.json +51 -0
- checkpoint-2772/tokenizer.json +3 -0
- checkpoint-2772/tokenizer_config.json +65 -0
- checkpoint-2772/trainer_state.json +142 -0
1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": false,
|
| 4 |
+
"pooling_mode_mean_tokens": true,
|
| 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 |
+
}
|
checkpoint-1386/1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": false,
|
| 4 |
+
"pooling_mode_mean_tokens": true,
|
| 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 |
+
}
|
checkpoint-1386/README.md
ADDED
|
@@ -0,0 +1,460 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- sentence-transformers
|
| 7 |
+
- sentence-similarity
|
| 8 |
+
- feature-extraction
|
| 9 |
+
- generated_from_trainer
|
| 10 |
+
- dataset_size:2130621
|
| 11 |
+
- loss:ContrastiveLoss
|
| 12 |
+
base_model: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
|
| 13 |
+
widget:
|
| 14 |
+
- source_sentence: Kim Chol-sam
|
| 15 |
+
sentences:
|
| 16 |
+
- Stankevich Sergey Nikolayevich
|
| 17 |
+
- Kim Chin-So’k
|
| 18 |
+
- Julen Lopetegui Agote
|
| 19 |
+
- source_sentence: دينا بنت عبد الحميد
|
| 20 |
+
sentences:
|
| 21 |
+
- Alexia van Amsberg
|
| 22 |
+
- Anthony Nicholas Colin Maitland Biddulph, 5th Baron Biddulph
|
| 23 |
+
- Dina bint Abdul-Hamíd
|
| 24 |
+
- source_sentence: Մուհամեդ բեն Նաիֆ Ալ Սաուդ
|
| 25 |
+
sentences:
|
| 26 |
+
- Karpov Anatoly Evgenyevich
|
| 27 |
+
- GNPower Mariveles Coal Plant [former]
|
| 28 |
+
- Muhammed bin Nayef bin Abdul Aziz Al Saud
|
| 29 |
+
- source_sentence: Edward Gnehm
|
| 30 |
+
sentences:
|
| 31 |
+
- Шауэрте, Хартмут
|
| 32 |
+
- Ханзада Филипп, Эдинбург герцогі
|
| 33 |
+
- AFX
|
| 34 |
+
- source_sentence: Schori i Lidingö
|
| 35 |
+
sentences:
|
| 36 |
+
- Yordan Canev
|
| 37 |
+
- ကားပေါ့ အန်နာတိုလီ
|
| 38 |
+
- BYSTROV, Mikhail Ivanovich
|
| 39 |
+
pipeline_tag: sentence-similarity
|
| 40 |
+
library_name: sentence-transformers
|
| 41 |
+
metrics:
|
| 42 |
+
- cosine_accuracy
|
| 43 |
+
- cosine_accuracy_threshold
|
| 44 |
+
- cosine_f1
|
| 45 |
+
- cosine_f1_threshold
|
| 46 |
+
- cosine_precision
|
| 47 |
+
- cosine_recall
|
| 48 |
+
- cosine_ap
|
| 49 |
+
- cosine_mcc
|
| 50 |
+
model-index:
|
| 51 |
+
- name: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2-name-matcher-original
|
| 52 |
+
results:
|
| 53 |
+
- task:
|
| 54 |
+
type: binary-classification
|
| 55 |
+
name: Binary Classification
|
| 56 |
+
dataset:
|
| 57 |
+
name: sentence transformers paraphrase multilingual MiniLM L12 v2
|
| 58 |
+
type: sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2
|
| 59 |
+
metrics:
|
| 60 |
+
- type: cosine_accuracy
|
| 61 |
+
value: 0.9846242227629088
|
| 62 |
+
name: Cosine Accuracy
|
| 63 |
+
- type: cosine_accuracy_threshold
|
| 64 |
+
value: 0.6801187992095947
|
| 65 |
+
name: Cosine Accuracy Threshold
|
| 66 |
+
- type: cosine_f1
|
| 67 |
+
value: 0.9765449140552956
|
| 68 |
+
name: Cosine F1
|
| 69 |
+
- type: cosine_f1_threshold
|
| 70 |
+
value: 0.6780189275741577
|
| 71 |
+
name: Cosine F1 Threshold
|
| 72 |
+
- type: cosine_precision
|
| 73 |
+
value: 0.9721848413657824
|
| 74 |
+
name: Cosine Precision
|
| 75 |
+
- type: cosine_recall
|
| 76 |
+
value: 0.9809442711989229
|
| 77 |
+
name: Cosine Recall
|
| 78 |
+
- type: cosine_ap
|
| 79 |
+
value: 0.9955904030209028
|
| 80 |
+
name: Cosine Ap
|
| 81 |
+
- type: cosine_mcc
|
| 82 |
+
value: 0.9651303277408154
|
| 83 |
+
name: Cosine Mcc
|
| 84 |
+
---
|
| 85 |
+
|
| 86 |
+
# sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2-name-matcher-original
|
| 87 |
+
|
| 88 |
+
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2). It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
|
| 89 |
+
|
| 90 |
+
## Model Details
|
| 91 |
+
|
| 92 |
+
### Model Description
|
| 93 |
+
- **Model Type:** Sentence Transformer
|
| 94 |
+
- **Base model:** [sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2) <!-- at revision 86741b4e3f5cb7765a600d3a3d55a0f6a6cb443d -->
|
| 95 |
+
- **Maximum Sequence Length:** 128 tokens
|
| 96 |
+
- **Output Dimensionality:** 384 dimensions
|
| 97 |
+
- **Similarity Function:** Cosine Similarity
|
| 98 |
+
<!-- - **Training Dataset:** Unknown -->
|
| 99 |
+
- **Language:** en
|
| 100 |
+
- **License:** apache-2.0
|
| 101 |
+
|
| 102 |
+
### Model Sources
|
| 103 |
+
|
| 104 |
+
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
|
| 105 |
+
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
|
| 106 |
+
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
|
| 107 |
+
|
| 108 |
+
### Full Model Architecture
|
| 109 |
+
|
| 110 |
+
```
|
| 111 |
+
SentenceTransformer(
|
| 112 |
+
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
|
| 113 |
+
(1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
|
| 114 |
+
)
|
| 115 |
+
```
|
| 116 |
+
|
| 117 |
+
## Usage
|
| 118 |
+
|
| 119 |
+
### Direct Usage (Sentence Transformers)
|
| 120 |
+
|
| 121 |
+
First install the Sentence Transformers library:
|
| 122 |
+
|
| 123 |
+
```bash
|
| 124 |
+
pip install -U sentence-transformers
|
| 125 |
+
```
|
| 126 |
+
|
| 127 |
+
Then you can load this model and run inference.
|
| 128 |
+
```python
|
| 129 |
+
from sentence_transformers import SentenceTransformer
|
| 130 |
+
|
| 131 |
+
# Download from the 🤗 Hub
|
| 132 |
+
model = SentenceTransformer("sentence_transformers_model_id")
|
| 133 |
+
# Run inference
|
| 134 |
+
sentences = [
|
| 135 |
+
'Schori i Lidingö',
|
| 136 |
+
'Yordan Canev',
|
| 137 |
+
'ကားပေါ့ အန်နာတိုလီ',
|
| 138 |
+
]
|
| 139 |
+
embeddings = model.encode(sentences)
|
| 140 |
+
print(embeddings.shape)
|
| 141 |
+
# [3, 384]
|
| 142 |
+
|
| 143 |
+
# Get the similarity scores for the embeddings
|
| 144 |
+
similarities = model.similarity(embeddings, embeddings)
|
| 145 |
+
print(similarities.shape)
|
| 146 |
+
# [3, 3]
|
| 147 |
+
```
|
| 148 |
+
|
| 149 |
+
<!--
|
| 150 |
+
### Direct Usage (Transformers)
|
| 151 |
+
|
| 152 |
+
<details><summary>Click to see the direct usage in Transformers</summary>
|
| 153 |
+
|
| 154 |
+
</details>
|
| 155 |
+
-->
|
| 156 |
+
|
| 157 |
+
<!--
|
| 158 |
+
### Downstream Usage (Sentence Transformers)
|
| 159 |
+
|
| 160 |
+
You can finetune this model on your own dataset.
|
| 161 |
+
|
| 162 |
+
<details><summary>Click to expand</summary>
|
| 163 |
+
|
| 164 |
+
</details>
|
| 165 |
+
-->
|
| 166 |
+
|
| 167 |
+
<!--
|
| 168 |
+
### Out-of-Scope Use
|
| 169 |
+
|
| 170 |
+
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
|
| 171 |
+
-->
|
| 172 |
+
|
| 173 |
+
## Evaluation
|
| 174 |
+
|
| 175 |
+
### Metrics
|
| 176 |
+
|
| 177 |
+
#### Binary Classification
|
| 178 |
+
|
| 179 |
+
* Dataset: `sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2`
|
| 180 |
+
* Evaluated with [<code>BinaryClassificationEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.BinaryClassificationEvaluator)
|
| 181 |
+
|
| 182 |
+
| Metric | Value |
|
| 183 |
+
|:--------------------------|:-----------|
|
| 184 |
+
| cosine_accuracy | 0.9846 |
|
| 185 |
+
| cosine_accuracy_threshold | 0.6801 |
|
| 186 |
+
| cosine_f1 | 0.9765 |
|
| 187 |
+
| cosine_f1_threshold | 0.678 |
|
| 188 |
+
| cosine_precision | 0.9722 |
|
| 189 |
+
| cosine_recall | 0.9809 |
|
| 190 |
+
| **cosine_ap** | **0.9956** |
|
| 191 |
+
| cosine_mcc | 0.9651 |
|
| 192 |
+
|
| 193 |
+
<!--
|
| 194 |
+
## Bias, Risks and Limitations
|
| 195 |
+
|
| 196 |
+
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
|
| 197 |
+
-->
|
| 198 |
+
|
| 199 |
+
<!--
|
| 200 |
+
### Recommendations
|
| 201 |
+
|
| 202 |
+
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
|
| 203 |
+
-->
|
| 204 |
+
|
| 205 |
+
## Training Details
|
| 206 |
+
|
| 207 |
+
### Training Dataset
|
| 208 |
+
|
| 209 |
+
#### Unnamed Dataset
|
| 210 |
+
|
| 211 |
+
* Size: 2,130,621 training samples
|
| 212 |
+
* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
|
| 213 |
+
* Approximate statistics based on the first 1000 samples:
|
| 214 |
+
| | sentence1 | sentence2 | label |
|
| 215 |
+
|:--------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------|
|
| 216 |
+
| type | string | string | float |
|
| 217 |
+
| details | <ul><li>min: 3 tokens</li><li>mean: 9.32 tokens</li><li>max: 57 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 9.16 tokens</li><li>max: 54 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.34</li><li>max: 1.0</li></ul> |
|
| 218 |
+
* Samples:
|
| 219 |
+
| sentence1 | sentence2 | label |
|
| 220 |
+
|:----------------------------------|:------------------------------------|:-----------------|
|
| 221 |
+
| <code>캐스린 설리번</code> | <code>Kathryn D. Sullivanová</code> | <code>1.0</code> |
|
| 222 |
+
| <code>ଶିବରାଜ ଅଧାଲରାଓ ପାଟିଲ</code> | <code>Aleksander Lubocki</code> | <code>0.0</code> |
|
| 223 |
+
| <code>Пырванов, Георги</code> | <code>アナトーリー・セルジュコフ</code> | <code>0.0</code> |
|
| 224 |
+
* Loss: [<code>ContrastiveLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastiveloss) with these parameters:
|
| 225 |
+
```json
|
| 226 |
+
{
|
| 227 |
+
"distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
|
| 228 |
+
"margin": 0.5,
|
| 229 |
+
"size_average": true
|
| 230 |
+
}
|
| 231 |
+
```
|
| 232 |
+
|
| 233 |
+
### Evaluation Dataset
|
| 234 |
+
|
| 235 |
+
#### Unnamed Dataset
|
| 236 |
+
|
| 237 |
+
* Size: 2,663,276 evaluation samples
|
| 238 |
+
* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
|
| 239 |
+
* Approximate statistics based on the first 1000 samples:
|
| 240 |
+
| | sentence1 | sentence2 | label |
|
| 241 |
+
|:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
|
| 242 |
+
| type | string | string | float |
|
| 243 |
+
| details | <ul><li>min: 3 tokens</li><li>mean: 9.34 tokens</li><li>max: 102 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 9.11 tokens</li><li>max: 100 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.33</li><li>max: 1.0</li></ul> |
|
| 244 |
+
* Samples:
|
| 245 |
+
| sentence1 | sentence2 | label |
|
| 246 |
+
|:--------------------------------------|:---------------------------------------|:-----------------|
|
| 247 |
+
| <code>Ева Херман</code> | <code>I Xuan Karlos</code> | <code>0.0</code> |
|
| 248 |
+
| <code>Кличков Андрій Євгенович</code> | <code>Андрэй Яўгенавіч Клычкоў</code> | <code>1.0</code> |
|
| 249 |
+
| <code>Кинах А.</code> | <code>Senator John Hickenlooper</code> | <code>0.0</code> |
|
| 250 |
+
* Loss: [<code>ContrastiveLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastiveloss) with these parameters:
|
| 251 |
+
```json
|
| 252 |
+
{
|
| 253 |
+
"distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
|
| 254 |
+
"margin": 0.5,
|
| 255 |
+
"size_average": true
|
| 256 |
+
}
|
| 257 |
+
```
|
| 258 |
+
|
| 259 |
+
### Training Hyperparameters
|
| 260 |
+
#### Non-Default Hyperparameters
|
| 261 |
+
|
| 262 |
+
- `eval_strategy`: epoch
|
| 263 |
+
- `per_device_train_batch_size`: 768
|
| 264 |
+
- `per_device_eval_batch_size`: 768
|
| 265 |
+
- `gradient_accumulation_steps`: 4
|
| 266 |
+
- `learning_rate`: 3e-05
|
| 267 |
+
- `weight_decay`: 0.01
|
| 268 |
+
- `num_train_epochs`: 4
|
| 269 |
+
- `warmup_ratio`: 0.1
|
| 270 |
+
- `fp16`: True
|
| 271 |
+
- `load_best_model_at_end`: True
|
| 272 |
+
- `optim`: adafactor
|
| 273 |
+
|
| 274 |
+
#### All Hyperparameters
|
| 275 |
+
<details><summary>Click to expand</summary>
|
| 276 |
+
|
| 277 |
+
- `overwrite_output_dir`: False
|
| 278 |
+
- `do_predict`: False
|
| 279 |
+
- `eval_strategy`: epoch
|
| 280 |
+
- `prediction_loss_only`: True
|
| 281 |
+
- `per_device_train_batch_size`: 768
|
| 282 |
+
- `per_device_eval_batch_size`: 768
|
| 283 |
+
- `per_gpu_train_batch_size`: None
|
| 284 |
+
- `per_gpu_eval_batch_size`: None
|
| 285 |
+
- `gradient_accumulation_steps`: 4
|
| 286 |
+
- `eval_accumulation_steps`: None
|
| 287 |
+
- `torch_empty_cache_steps`: None
|
| 288 |
+
- `learning_rate`: 3e-05
|
| 289 |
+
- `weight_decay`: 0.01
|
| 290 |
+
- `adam_beta1`: 0.9
|
| 291 |
+
- `adam_beta2`: 0.999
|
| 292 |
+
- `adam_epsilon`: 1e-08
|
| 293 |
+
- `max_grad_norm`: 1.0
|
| 294 |
+
- `num_train_epochs`: 4
|
| 295 |
+
- `max_steps`: -1
|
| 296 |
+
- `lr_scheduler_type`: linear
|
| 297 |
+
- `lr_scheduler_kwargs`: {}
|
| 298 |
+
- `warmup_ratio`: 0.1
|
| 299 |
+
- `warmup_steps`: 0
|
| 300 |
+
- `log_level`: passive
|
| 301 |
+
- `log_level_replica`: warning
|
| 302 |
+
- `log_on_each_node`: True
|
| 303 |
+
- `logging_nan_inf_filter`: True
|
| 304 |
+
- `save_safetensors`: True
|
| 305 |
+
- `save_on_each_node`: False
|
| 306 |
+
- `save_only_model`: False
|
| 307 |
+
- `restore_callback_states_from_checkpoint`: False
|
| 308 |
+
- `no_cuda`: False
|
| 309 |
+
- `use_cpu`: False
|
| 310 |
+
- `use_mps_device`: False
|
| 311 |
+
- `seed`: 42
|
| 312 |
+
- `data_seed`: None
|
| 313 |
+
- `jit_mode_eval`: False
|
| 314 |
+
- `use_ipex`: False
|
| 315 |
+
- `bf16`: False
|
| 316 |
+
- `fp16`: True
|
| 317 |
+
- `fp16_opt_level`: O1
|
| 318 |
+
- `half_precision_backend`: auto
|
| 319 |
+
- `bf16_full_eval`: False
|
| 320 |
+
- `fp16_full_eval`: False
|
| 321 |
+
- `tf32`: None
|
| 322 |
+
- `local_rank`: 0
|
| 323 |
+
- `ddp_backend`: None
|
| 324 |
+
- `tpu_num_cores`: None
|
| 325 |
+
- `tpu_metrics_debug`: False
|
| 326 |
+
- `debug`: []
|
| 327 |
+
- `dataloader_drop_last`: False
|
| 328 |
+
- `dataloader_num_workers`: 0
|
| 329 |
+
- `dataloader_prefetch_factor`: None
|
| 330 |
+
- `past_index`: -1
|
| 331 |
+
- `disable_tqdm`: False
|
| 332 |
+
- `remove_unused_columns`: True
|
| 333 |
+
- `label_names`: None
|
| 334 |
+
- `load_best_model_at_end`: True
|
| 335 |
+
- `ignore_data_skip`: False
|
| 336 |
+
- `fsdp`: []
|
| 337 |
+
- `fsdp_min_num_params`: 0
|
| 338 |
+
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
|
| 339 |
+
- `tp_size`: 0
|
| 340 |
+
- `fsdp_transformer_layer_cls_to_wrap`: None
|
| 341 |
+
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
|
| 342 |
+
- `deepspeed`: None
|
| 343 |
+
- `label_smoothing_factor`: 0.0
|
| 344 |
+
- `optim`: adafactor
|
| 345 |
+
- `optim_args`: None
|
| 346 |
+
- `adafactor`: False
|
| 347 |
+
- `group_by_length`: False
|
| 348 |
+
- `length_column_name`: length
|
| 349 |
+
- `ddp_find_unused_parameters`: None
|
| 350 |
+
- `ddp_bucket_cap_mb`: None
|
| 351 |
+
- `ddp_broadcast_buffers`: False
|
| 352 |
+
- `dataloader_pin_memory`: True
|
| 353 |
+
- `dataloader_persistent_workers`: False
|
| 354 |
+
- `skip_memory_metrics`: True
|
| 355 |
+
- `use_legacy_prediction_loop`: False
|
| 356 |
+
- `push_to_hub`: False
|
| 357 |
+
- `resume_from_checkpoint`: None
|
| 358 |
+
- `hub_model_id`: None
|
| 359 |
+
- `hub_strategy`: every_save
|
| 360 |
+
- `hub_private_repo`: None
|
| 361 |
+
- `hub_always_push`: False
|
| 362 |
+
- `gradient_checkpointing`: False
|
| 363 |
+
- `gradient_checkpointing_kwargs`: None
|
| 364 |
+
- `include_inputs_for_metrics`: False
|
| 365 |
+
- `include_for_metrics`: []
|
| 366 |
+
- `eval_do_concat_batches`: True
|
| 367 |
+
- `fp16_backend`: auto
|
| 368 |
+
- `push_to_hub_model_id`: None
|
| 369 |
+
- `push_to_hub_organization`: None
|
| 370 |
+
- `mp_parameters`:
|
| 371 |
+
- `auto_find_batch_size`: False
|
| 372 |
+
- `full_determinism`: False
|
| 373 |
+
- `torchdynamo`: None
|
| 374 |
+
- `ray_scope`: last
|
| 375 |
+
- `ddp_timeout`: 1800
|
| 376 |
+
- `torch_compile`: False
|
| 377 |
+
- `torch_compile_backend`: None
|
| 378 |
+
- `torch_compile_mode`: None
|
| 379 |
+
- `include_tokens_per_second`: False
|
| 380 |
+
- `include_num_input_tokens_seen`: False
|
| 381 |
+
- `neftune_noise_alpha`: None
|
| 382 |
+
- `optim_target_modules`: None
|
| 383 |
+
- `batch_eval_metrics`: False
|
| 384 |
+
- `eval_on_start`: False
|
| 385 |
+
- `use_liger_kernel`: False
|
| 386 |
+
- `eval_use_gather_object`: False
|
| 387 |
+
- `average_tokens_across_devices`: False
|
| 388 |
+
- `prompts`: None
|
| 389 |
+
- `batch_sampler`: batch_sampler
|
| 390 |
+
- `multi_dataset_batch_sampler`: proportional
|
| 391 |
+
|
| 392 |
+
</details>
|
| 393 |
+
|
| 394 |
+
### Training Logs
|
| 395 |
+
| Epoch | Step | Training Loss | Validation Loss | sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap |
|
| 396 |
+
|:------:|:----:|:-------------:|:---------------:|:---------------------------------------------------------------------:|
|
| 397 |
+
| -1 | -1 | - | - | 0.7140 |
|
| 398 |
+
| 0.7207 | 500 | 0.038 | - | - |
|
| 399 |
+
| 0.9989 | 693 | - | 0.0028 | 0.9911 |
|
| 400 |
+
| 1.4425 | 1000 | 0.0128 | - | - |
|
| 401 |
+
| 1.9989 | 1386 | - | 0.0021 | 0.9956 |
|
| 402 |
+
|
| 403 |
+
|
| 404 |
+
### Framework Versions
|
| 405 |
+
- Python: 3.12.9
|
| 406 |
+
- Sentence Transformers: 3.4.1
|
| 407 |
+
- Transformers: 4.51.3
|
| 408 |
+
- PyTorch: 2.7.0+cu126
|
| 409 |
+
- Accelerate: 1.6.0
|
| 410 |
+
- Datasets: 3.6.0
|
| 411 |
+
- Tokenizers: 0.21.1
|
| 412 |
+
|
| 413 |
+
## Citation
|
| 414 |
+
|
| 415 |
+
### BibTeX
|
| 416 |
+
|
| 417 |
+
#### Sentence Transformers
|
| 418 |
+
```bibtex
|
| 419 |
+
@inproceedings{reimers-2019-sentence-bert,
|
| 420 |
+
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
|
| 421 |
+
author = "Reimers, Nils and Gurevych, Iryna",
|
| 422 |
+
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
|
| 423 |
+
month = "11",
|
| 424 |
+
year = "2019",
|
| 425 |
+
publisher = "Association for Computational Linguistics",
|
| 426 |
+
url = "https://arxiv.org/abs/1908.10084",
|
| 427 |
+
}
|
| 428 |
+
```
|
| 429 |
+
|
| 430 |
+
#### ContrastiveLoss
|
| 431 |
+
```bibtex
|
| 432 |
+
@inproceedings{hadsell2006dimensionality,
|
| 433 |
+
author={Hadsell, R. and Chopra, S. and LeCun, Y.},
|
| 434 |
+
booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)},
|
| 435 |
+
title={Dimensionality Reduction by Learning an Invariant Mapping},
|
| 436 |
+
year={2006},
|
| 437 |
+
volume={2},
|
| 438 |
+
number={},
|
| 439 |
+
pages={1735-1742},
|
| 440 |
+
doi={10.1109/CVPR.2006.100}
|
| 441 |
+
}
|
| 442 |
+
```
|
| 443 |
+
|
| 444 |
+
<!--
|
| 445 |
+
## Glossary
|
| 446 |
+
|
| 447 |
+
*Clearly define terms in order to be accessible across audiences.*
|
| 448 |
+
-->
|
| 449 |
+
|
| 450 |
+
<!--
|
| 451 |
+
## Model Card Authors
|
| 452 |
+
|
| 453 |
+
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
|
| 454 |
+
-->
|
| 455 |
+
|
| 456 |
+
<!--
|
| 457 |
+
## Model Card Contact
|
| 458 |
+
|
| 459 |
+
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
|
| 460 |
+
-->
|
checkpoint-1386/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertModel"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"gradient_checkpointing": false,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 384,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 1536,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.51.3",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 250037
|
| 25 |
+
}
|
checkpoint-1386/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.4.1",
|
| 4 |
+
"transformers": "4.51.3",
|
| 5 |
+
"pytorch": "2.7.0+cu126"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": "cosine"
|
| 10 |
+
}
|
checkpoint-1386/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e63c4102bbc807f6f65e6dfc340ebdf81b6d809419616e5c686dc0ae4ee24c69
|
| 3 |
+
size 470637416
|
checkpoint-1386/modules.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
]
|
checkpoint-1386/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:23860257dbcfd9b5b597ed0f71ec9d1d0888d586ddccbe366e3c4cea28a49c54
|
| 3 |
+
size 1715019
|
checkpoint-1386/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cafafda98c3eee33da2fc3dc512fb690207c63cd6c23669bf103b6115333673f
|
| 3 |
+
size 14645
|
checkpoint-1386/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:80cf4f866900877d5a70b5a4212850bd6349868cb41fa3deed374b2e1ff32fd1
|
| 3 |
+
size 1383
|
checkpoint-1386/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cf0f4146f2cb078d3608133188b68bec6910432c8f56446c257e96b3d46d2b17
|
| 3 |
+
size 1465
|
checkpoint-1386/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 128,
|
| 3 |
+
"do_lower_case": false
|
| 4 |
+
}
|
checkpoint-1386/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
checkpoint-1386/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cad551d5600a84242d0973327029452a1e3672ba6313c2a3c3d69c4310e12719
|
| 3 |
+
size 17082987
|
checkpoint-1386/tokenizer_config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"250001": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"eos_token": "</s>",
|
| 49 |
+
"extra_special_tokens": {},
|
| 50 |
+
"mask_token": "<mask>",
|
| 51 |
+
"max_length": 128,
|
| 52 |
+
"model_max_length": 128,
|
| 53 |
+
"pad_to_multiple_of": null,
|
| 54 |
+
"pad_token": "<pad>",
|
| 55 |
+
"pad_token_type_id": 0,
|
| 56 |
+
"padding_side": "right",
|
| 57 |
+
"sep_token": "</s>",
|
| 58 |
+
"stride": 0,
|
| 59 |
+
"strip_accents": null,
|
| 60 |
+
"tokenize_chinese_chars": true,
|
| 61 |
+
"tokenizer_class": "BertTokenizer",
|
| 62 |
+
"truncation_side": "right",
|
| 63 |
+
"truncation_strategy": "longest_first",
|
| 64 |
+
"unk_token": "<unk>"
|
| 65 |
+
}
|
checkpoint-1386/trainer_state.json
ADDED
|
@@ -0,0 +1,89 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 1386,
|
| 3 |
+
"best_metric": 0.002118302509188652,
|
| 4 |
+
"best_model_checkpoint": "data/fine-tuned-sbert-sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2-original-adafactor/checkpoint-1386",
|
| 5 |
+
"epoch": 1.998918918918919,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 1386,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.7207207207207207,
|
| 14 |
+
"grad_norm": 0.15965162217617035,
|
| 15 |
+
"learning_rate": 2.7341619887730554e-05,
|
| 16 |
+
"loss": 0.038,
|
| 17 |
+
"step": 500
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.9989189189189189,
|
| 21 |
+
"eval_loss": 0.002777691464871168,
|
| 22 |
+
"eval_runtime": 792.7494,
|
| 23 |
+
"eval_samples_per_second": 3359.543,
|
| 24 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.975327415818089,
|
| 25 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.7701693773269653,
|
| 26 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.9911172257655966,
|
| 27 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9624723810391563,
|
| 28 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.7597355842590332,
|
| 29 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9441287064717891,
|
| 30 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9527175567355498,
|
| 31 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.9724290300680068,
|
| 32 |
+
"eval_steps_per_second": 4.375,
|
| 33 |
+
"step": 693
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 1.4425225225225224,
|
| 37 |
+
"grad_norm": 0.11966603249311447,
|
| 38 |
+
"learning_rate": 2.1327185244587008e-05,
|
| 39 |
+
"loss": 0.0128,
|
| 40 |
+
"step": 1000
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 1.998918918918919,
|
| 44 |
+
"eval_loss": 0.002118302509188652,
|
| 45 |
+
"eval_runtime": 789.0568,
|
| 46 |
+
"eval_samples_per_second": 3375.265,
|
| 47 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.9846242227629088,
|
| 48 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.6801187992095947,
|
| 49 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.9955904030209028,
|
| 50 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9765449140552956,
|
| 51 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.6780189275741577,
|
| 52 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9651303277408154,
|
| 53 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9721848413657824,
|
| 54 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.9809442711989229,
|
| 55 |
+
"eval_steps_per_second": 4.395,
|
| 56 |
+
"step": 1386
|
| 57 |
+
}
|
| 58 |
+
],
|
| 59 |
+
"logging_steps": 500,
|
| 60 |
+
"max_steps": 2772,
|
| 61 |
+
"num_input_tokens_seen": 0,
|
| 62 |
+
"num_train_epochs": 4,
|
| 63 |
+
"save_steps": 100,
|
| 64 |
+
"stateful_callbacks": {
|
| 65 |
+
"EarlyStoppingCallback": {
|
| 66 |
+
"args": {
|
| 67 |
+
"early_stopping_patience": 1,
|
| 68 |
+
"early_stopping_threshold": 0.0
|
| 69 |
+
},
|
| 70 |
+
"attributes": {
|
| 71 |
+
"early_stopping_patience_counter": 0
|
| 72 |
+
}
|
| 73 |
+
},
|
| 74 |
+
"TrainerControl": {
|
| 75 |
+
"args": {
|
| 76 |
+
"should_epoch_stop": false,
|
| 77 |
+
"should_evaluate": false,
|
| 78 |
+
"should_log": false,
|
| 79 |
+
"should_save": true,
|
| 80 |
+
"should_training_stop": false
|
| 81 |
+
},
|
| 82 |
+
"attributes": {}
|
| 83 |
+
}
|
| 84 |
+
},
|
| 85 |
+
"total_flos": 0.0,
|
| 86 |
+
"train_batch_size": 768,
|
| 87 |
+
"trial_name": null,
|
| 88 |
+
"trial_params": null
|
| 89 |
+
}
|
checkpoint-1386/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee490aac80277022fdde76dd6763dbabee88802149f3f246e7d0572b504fcdaf
|
| 3 |
+
size 6097
|
checkpoint-1386/unigram.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da145b5e7700ae40f16691ec32a0b1fdc1ee3298db22a31ea55f57a966c4a65d
|
| 3 |
+
size 14763260
|
checkpoint-2079/1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": false,
|
| 4 |
+
"pooling_mode_mean_tokens": true,
|
| 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 |
+
}
|
checkpoint-2079/README.md
ADDED
|
@@ -0,0 +1,463 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- sentence-transformers
|
| 7 |
+
- sentence-similarity
|
| 8 |
+
- feature-extraction
|
| 9 |
+
- generated_from_trainer
|
| 10 |
+
- dataset_size:2130621
|
| 11 |
+
- loss:ContrastiveLoss
|
| 12 |
+
base_model: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
|
| 13 |
+
widget:
|
| 14 |
+
- source_sentence: Kim Chol-sam
|
| 15 |
+
sentences:
|
| 16 |
+
- Stankevich Sergey Nikolayevich
|
| 17 |
+
- Kim Chin-So’k
|
| 18 |
+
- Julen Lopetegui Agote
|
| 19 |
+
- source_sentence: دينا بنت عبد الحميد
|
| 20 |
+
sentences:
|
| 21 |
+
- Alexia van Amsberg
|
| 22 |
+
- Anthony Nicholas Colin Maitland Biddulph, 5th Baron Biddulph
|
| 23 |
+
- Dina bint Abdul-Hamíd
|
| 24 |
+
- source_sentence: Մուհամեդ բեն Նաիֆ Ալ Սաուդ
|
| 25 |
+
sentences:
|
| 26 |
+
- Karpov Anatoly Evgenyevich
|
| 27 |
+
- GNPower Mariveles Coal Plant [former]
|
| 28 |
+
- Muhammed bin Nayef bin Abdul Aziz Al Saud
|
| 29 |
+
- source_sentence: Edward Gnehm
|
| 30 |
+
sentences:
|
| 31 |
+
- Шауэрте, Хартмут
|
| 32 |
+
- Ханзада Филипп, Эдинбург герцогі
|
| 33 |
+
- AFX
|
| 34 |
+
- source_sentence: Schori i Lidingö
|
| 35 |
+
sentences:
|
| 36 |
+
- Yordan Canev
|
| 37 |
+
- ကားပေါ့ အန်နာတိုလီ
|
| 38 |
+
- BYSTROV, Mikhail Ivanovich
|
| 39 |
+
pipeline_tag: sentence-similarity
|
| 40 |
+
library_name: sentence-transformers
|
| 41 |
+
metrics:
|
| 42 |
+
- cosine_accuracy
|
| 43 |
+
- cosine_accuracy_threshold
|
| 44 |
+
- cosine_f1
|
| 45 |
+
- cosine_f1_threshold
|
| 46 |
+
- cosine_precision
|
| 47 |
+
- cosine_recall
|
| 48 |
+
- cosine_ap
|
| 49 |
+
- cosine_mcc
|
| 50 |
+
model-index:
|
| 51 |
+
- name: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2-name-matcher-original
|
| 52 |
+
results:
|
| 53 |
+
- task:
|
| 54 |
+
type: binary-classification
|
| 55 |
+
name: Binary Classification
|
| 56 |
+
dataset:
|
| 57 |
+
name: sentence transformers paraphrase multilingual MiniLM L12 v2
|
| 58 |
+
type: sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2
|
| 59 |
+
metrics:
|
| 60 |
+
- type: cosine_accuracy
|
| 61 |
+
value: 0.9879171547865789
|
| 62 |
+
name: Cosine Accuracy
|
| 63 |
+
- type: cosine_accuracy_threshold
|
| 64 |
+
value: 0.7181636691093445
|
| 65 |
+
name: Cosine Accuracy Threshold
|
| 66 |
+
- type: cosine_f1
|
| 67 |
+
value: 0.9815604299892273
|
| 68 |
+
name: Cosine F1
|
| 69 |
+
- type: cosine_f1_threshold
|
| 70 |
+
value: 0.7181636691093445
|
| 71 |
+
name: Cosine F1 Threshold
|
| 72 |
+
- type: cosine_precision
|
| 73 |
+
value: 0.9775832353646149
|
| 74 |
+
name: Cosine Precision
|
| 75 |
+
- type: cosine_recall
|
| 76 |
+
value: 0.98557011840788
|
| 77 |
+
name: Cosine Recall
|
| 78 |
+
- type: cosine_ap
|
| 79 |
+
value: 0.996840725826042
|
| 80 |
+
name: Cosine Ap
|
| 81 |
+
- type: cosine_mcc
|
| 82 |
+
value: 0.9725931427811844
|
| 83 |
+
name: Cosine Mcc
|
| 84 |
+
---
|
| 85 |
+
|
| 86 |
+
# sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2-name-matcher-original
|
| 87 |
+
|
| 88 |
+
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2). It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
|
| 89 |
+
|
| 90 |
+
## Model Details
|
| 91 |
+
|
| 92 |
+
### Model Description
|
| 93 |
+
- **Model Type:** Sentence Transformer
|
| 94 |
+
- **Base model:** [sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2) <!-- at revision 86741b4e3f5cb7765a600d3a3d55a0f6a6cb443d -->
|
| 95 |
+
- **Maximum Sequence Length:** 128 tokens
|
| 96 |
+
- **Output Dimensionality:** 384 dimensions
|
| 97 |
+
- **Similarity Function:** Cosine Similarity
|
| 98 |
+
<!-- - **Training Dataset:** Unknown -->
|
| 99 |
+
- **Language:** en
|
| 100 |
+
- **License:** apache-2.0
|
| 101 |
+
|
| 102 |
+
### Model Sources
|
| 103 |
+
|
| 104 |
+
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
|
| 105 |
+
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
|
| 106 |
+
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
|
| 107 |
+
|
| 108 |
+
### Full Model Architecture
|
| 109 |
+
|
| 110 |
+
```
|
| 111 |
+
SentenceTransformer(
|
| 112 |
+
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
|
| 113 |
+
(1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
|
| 114 |
+
)
|
| 115 |
+
```
|
| 116 |
+
|
| 117 |
+
## Usage
|
| 118 |
+
|
| 119 |
+
### Direct Usage (Sentence Transformers)
|
| 120 |
+
|
| 121 |
+
First install the Sentence Transformers library:
|
| 122 |
+
|
| 123 |
+
```bash
|
| 124 |
+
pip install -U sentence-transformers
|
| 125 |
+
```
|
| 126 |
+
|
| 127 |
+
Then you can load this model and run inference.
|
| 128 |
+
```python
|
| 129 |
+
from sentence_transformers import SentenceTransformer
|
| 130 |
+
|
| 131 |
+
# Download from the 🤗 Hub
|
| 132 |
+
model = SentenceTransformer("sentence_transformers_model_id")
|
| 133 |
+
# Run inference
|
| 134 |
+
sentences = [
|
| 135 |
+
'Schori i Lidingö',
|
| 136 |
+
'Yordan Canev',
|
| 137 |
+
'ကားပေါ့ အန်နာတိုလီ',
|
| 138 |
+
]
|
| 139 |
+
embeddings = model.encode(sentences)
|
| 140 |
+
print(embeddings.shape)
|
| 141 |
+
# [3, 384]
|
| 142 |
+
|
| 143 |
+
# Get the similarity scores for the embeddings
|
| 144 |
+
similarities = model.similarity(embeddings, embeddings)
|
| 145 |
+
print(similarities.shape)
|
| 146 |
+
# [3, 3]
|
| 147 |
+
```
|
| 148 |
+
|
| 149 |
+
<!--
|
| 150 |
+
### Direct Usage (Transformers)
|
| 151 |
+
|
| 152 |
+
<details><summary>Click to see the direct usage in Transformers</summary>
|
| 153 |
+
|
| 154 |
+
</details>
|
| 155 |
+
-->
|
| 156 |
+
|
| 157 |
+
<!--
|
| 158 |
+
### Downstream Usage (Sentence Transformers)
|
| 159 |
+
|
| 160 |
+
You can finetune this model on your own dataset.
|
| 161 |
+
|
| 162 |
+
<details><summary>Click to expand</summary>
|
| 163 |
+
|
| 164 |
+
</details>
|
| 165 |
+
-->
|
| 166 |
+
|
| 167 |
+
<!--
|
| 168 |
+
### Out-of-Scope Use
|
| 169 |
+
|
| 170 |
+
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
|
| 171 |
+
-->
|
| 172 |
+
|
| 173 |
+
## Evaluation
|
| 174 |
+
|
| 175 |
+
### Metrics
|
| 176 |
+
|
| 177 |
+
#### Binary Classification
|
| 178 |
+
|
| 179 |
+
* Dataset: `sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2`
|
| 180 |
+
* Evaluated with [<code>BinaryClassificationEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.BinaryClassificationEvaluator)
|
| 181 |
+
|
| 182 |
+
| Metric | Value |
|
| 183 |
+
|:--------------------------|:-----------|
|
| 184 |
+
| cosine_accuracy | 0.9879 |
|
| 185 |
+
| cosine_accuracy_threshold | 0.7182 |
|
| 186 |
+
| cosine_f1 | 0.9816 |
|
| 187 |
+
| cosine_f1_threshold | 0.7182 |
|
| 188 |
+
| cosine_precision | 0.9776 |
|
| 189 |
+
| cosine_recall | 0.9856 |
|
| 190 |
+
| **cosine_ap** | **0.9968** |
|
| 191 |
+
| cosine_mcc | 0.9726 |
|
| 192 |
+
|
| 193 |
+
<!--
|
| 194 |
+
## Bias, Risks and Limitations
|
| 195 |
+
|
| 196 |
+
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
|
| 197 |
+
-->
|
| 198 |
+
|
| 199 |
+
<!--
|
| 200 |
+
### Recommendations
|
| 201 |
+
|
| 202 |
+
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
|
| 203 |
+
-->
|
| 204 |
+
|
| 205 |
+
## Training Details
|
| 206 |
+
|
| 207 |
+
### Training Dataset
|
| 208 |
+
|
| 209 |
+
#### Unnamed Dataset
|
| 210 |
+
|
| 211 |
+
* Size: 2,130,621 training samples
|
| 212 |
+
* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
|
| 213 |
+
* Approximate statistics based on the first 1000 samples:
|
| 214 |
+
| | sentence1 | sentence2 | label |
|
| 215 |
+
|:--------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------|
|
| 216 |
+
| type | string | string | float |
|
| 217 |
+
| details | <ul><li>min: 3 tokens</li><li>mean: 9.32 tokens</li><li>max: 57 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 9.16 tokens</li><li>max: 54 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.34</li><li>max: 1.0</li></ul> |
|
| 218 |
+
* Samples:
|
| 219 |
+
| sentence1 | sentence2 | label |
|
| 220 |
+
|:----------------------------------|:------------------------------------|:-----------------|
|
| 221 |
+
| <code>캐스린 설리번</code> | <code>Kathryn D. Sullivanová</code> | <code>1.0</code> |
|
| 222 |
+
| <code>ଶିବରାଜ ଅଧାଲରାଓ ପାଟିଲ</code> | <code>Aleksander Lubocki</code> | <code>0.0</code> |
|
| 223 |
+
| <code>Пырванов, Георги</code> | <code>アナトーリー・セルジュコフ</code> | <code>0.0</code> |
|
| 224 |
+
* Loss: [<code>ContrastiveLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastiveloss) with these parameters:
|
| 225 |
+
```json
|
| 226 |
+
{
|
| 227 |
+
"distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
|
| 228 |
+
"margin": 0.5,
|
| 229 |
+
"size_average": true
|
| 230 |
+
}
|
| 231 |
+
```
|
| 232 |
+
|
| 233 |
+
### Evaluation Dataset
|
| 234 |
+
|
| 235 |
+
#### Unnamed Dataset
|
| 236 |
+
|
| 237 |
+
* Size: 2,663,276 evaluation samples
|
| 238 |
+
* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
|
| 239 |
+
* Approximate statistics based on the first 1000 samples:
|
| 240 |
+
| | sentence1 | sentence2 | label |
|
| 241 |
+
|:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
|
| 242 |
+
| type | string | string | float |
|
| 243 |
+
| details | <ul><li>min: 3 tokens</li><li>mean: 9.34 tokens</li><li>max: 102 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 9.11 tokens</li><li>max: 100 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.33</li><li>max: 1.0</li></ul> |
|
| 244 |
+
* Samples:
|
| 245 |
+
| sentence1 | sentence2 | label |
|
| 246 |
+
|:--------------------------------------|:---------------------------------------|:-----------------|
|
| 247 |
+
| <code>Ева Херман</code> | <code>I Xuan Karlos</code> | <code>0.0</code> |
|
| 248 |
+
| <code>Кличков Андрій Євгенович</code> | <code>Андрэй Яўгенавіч Клычкоў</code> | <code>1.0</code> |
|
| 249 |
+
| <code>Кинах А.</code> | <code>Senator John Hickenlooper</code> | <code>0.0</code> |
|
| 250 |
+
* Loss: [<code>ContrastiveLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastiveloss) with these parameters:
|
| 251 |
+
```json
|
| 252 |
+
{
|
| 253 |
+
"distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
|
| 254 |
+
"margin": 0.5,
|
| 255 |
+
"size_average": true
|
| 256 |
+
}
|
| 257 |
+
```
|
| 258 |
+
|
| 259 |
+
### Training Hyperparameters
|
| 260 |
+
#### Non-Default Hyperparameters
|
| 261 |
+
|
| 262 |
+
- `eval_strategy`: epoch
|
| 263 |
+
- `per_device_train_batch_size`: 768
|
| 264 |
+
- `per_device_eval_batch_size`: 768
|
| 265 |
+
- `gradient_accumulation_steps`: 4
|
| 266 |
+
- `learning_rate`: 3e-05
|
| 267 |
+
- `weight_decay`: 0.01
|
| 268 |
+
- `num_train_epochs`: 4
|
| 269 |
+
- `warmup_ratio`: 0.1
|
| 270 |
+
- `fp16`: True
|
| 271 |
+
- `load_best_model_at_end`: True
|
| 272 |
+
- `optim`: adafactor
|
| 273 |
+
|
| 274 |
+
#### All Hyperparameters
|
| 275 |
+
<details><summary>Click to expand</summary>
|
| 276 |
+
|
| 277 |
+
- `overwrite_output_dir`: False
|
| 278 |
+
- `do_predict`: False
|
| 279 |
+
- `eval_strategy`: epoch
|
| 280 |
+
- `prediction_loss_only`: True
|
| 281 |
+
- `per_device_train_batch_size`: 768
|
| 282 |
+
- `per_device_eval_batch_size`: 768
|
| 283 |
+
- `per_gpu_train_batch_size`: None
|
| 284 |
+
- `per_gpu_eval_batch_size`: None
|
| 285 |
+
- `gradient_accumulation_steps`: 4
|
| 286 |
+
- `eval_accumulation_steps`: None
|
| 287 |
+
- `torch_empty_cache_steps`: None
|
| 288 |
+
- `learning_rate`: 3e-05
|
| 289 |
+
- `weight_decay`: 0.01
|
| 290 |
+
- `adam_beta1`: 0.9
|
| 291 |
+
- `adam_beta2`: 0.999
|
| 292 |
+
- `adam_epsilon`: 1e-08
|
| 293 |
+
- `max_grad_norm`: 1.0
|
| 294 |
+
- `num_train_epochs`: 4
|
| 295 |
+
- `max_steps`: -1
|
| 296 |
+
- `lr_scheduler_type`: linear
|
| 297 |
+
- `lr_scheduler_kwargs`: {}
|
| 298 |
+
- `warmup_ratio`: 0.1
|
| 299 |
+
- `warmup_steps`: 0
|
| 300 |
+
- `log_level`: passive
|
| 301 |
+
- `log_level_replica`: warning
|
| 302 |
+
- `log_on_each_node`: True
|
| 303 |
+
- `logging_nan_inf_filter`: True
|
| 304 |
+
- `save_safetensors`: True
|
| 305 |
+
- `save_on_each_node`: False
|
| 306 |
+
- `save_only_model`: False
|
| 307 |
+
- `restore_callback_states_from_checkpoint`: False
|
| 308 |
+
- `no_cuda`: False
|
| 309 |
+
- `use_cpu`: False
|
| 310 |
+
- `use_mps_device`: False
|
| 311 |
+
- `seed`: 42
|
| 312 |
+
- `data_seed`: None
|
| 313 |
+
- `jit_mode_eval`: False
|
| 314 |
+
- `use_ipex`: False
|
| 315 |
+
- `bf16`: False
|
| 316 |
+
- `fp16`: True
|
| 317 |
+
- `fp16_opt_level`: O1
|
| 318 |
+
- `half_precision_backend`: auto
|
| 319 |
+
- `bf16_full_eval`: False
|
| 320 |
+
- `fp16_full_eval`: False
|
| 321 |
+
- `tf32`: None
|
| 322 |
+
- `local_rank`: 0
|
| 323 |
+
- `ddp_backend`: None
|
| 324 |
+
- `tpu_num_cores`: None
|
| 325 |
+
- `tpu_metrics_debug`: False
|
| 326 |
+
- `debug`: []
|
| 327 |
+
- `dataloader_drop_last`: False
|
| 328 |
+
- `dataloader_num_workers`: 0
|
| 329 |
+
- `dataloader_prefetch_factor`: None
|
| 330 |
+
- `past_index`: -1
|
| 331 |
+
- `disable_tqdm`: False
|
| 332 |
+
- `remove_unused_columns`: True
|
| 333 |
+
- `label_names`: None
|
| 334 |
+
- `load_best_model_at_end`: True
|
| 335 |
+
- `ignore_data_skip`: False
|
| 336 |
+
- `fsdp`: []
|
| 337 |
+
- `fsdp_min_num_params`: 0
|
| 338 |
+
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
|
| 339 |
+
- `tp_size`: 0
|
| 340 |
+
- `fsdp_transformer_layer_cls_to_wrap`: None
|
| 341 |
+
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
|
| 342 |
+
- `deepspeed`: None
|
| 343 |
+
- `label_smoothing_factor`: 0.0
|
| 344 |
+
- `optim`: adafactor
|
| 345 |
+
- `optim_args`: None
|
| 346 |
+
- `adafactor`: False
|
| 347 |
+
- `group_by_length`: False
|
| 348 |
+
- `length_column_name`: length
|
| 349 |
+
- `ddp_find_unused_parameters`: None
|
| 350 |
+
- `ddp_bucket_cap_mb`: None
|
| 351 |
+
- `ddp_broadcast_buffers`: False
|
| 352 |
+
- `dataloader_pin_memory`: True
|
| 353 |
+
- `dataloader_persistent_workers`: False
|
| 354 |
+
- `skip_memory_metrics`: True
|
| 355 |
+
- `use_legacy_prediction_loop`: False
|
| 356 |
+
- `push_to_hub`: False
|
| 357 |
+
- `resume_from_checkpoint`: None
|
| 358 |
+
- `hub_model_id`: None
|
| 359 |
+
- `hub_strategy`: every_save
|
| 360 |
+
- `hub_private_repo`: None
|
| 361 |
+
- `hub_always_push`: False
|
| 362 |
+
- `gradient_checkpointing`: False
|
| 363 |
+
- `gradient_checkpointing_kwargs`: None
|
| 364 |
+
- `include_inputs_for_metrics`: False
|
| 365 |
+
- `include_for_metrics`: []
|
| 366 |
+
- `eval_do_concat_batches`: True
|
| 367 |
+
- `fp16_backend`: auto
|
| 368 |
+
- `push_to_hub_model_id`: None
|
| 369 |
+
- `push_to_hub_organization`: None
|
| 370 |
+
- `mp_parameters`:
|
| 371 |
+
- `auto_find_batch_size`: False
|
| 372 |
+
- `full_determinism`: False
|
| 373 |
+
- `torchdynamo`: None
|
| 374 |
+
- `ray_scope`: last
|
| 375 |
+
- `ddp_timeout`: 1800
|
| 376 |
+
- `torch_compile`: False
|
| 377 |
+
- `torch_compile_backend`: None
|
| 378 |
+
- `torch_compile_mode`: None
|
| 379 |
+
- `include_tokens_per_second`: False
|
| 380 |
+
- `include_num_input_tokens_seen`: False
|
| 381 |
+
- `neftune_noise_alpha`: None
|
| 382 |
+
- `optim_target_modules`: None
|
| 383 |
+
- `batch_eval_metrics`: False
|
| 384 |
+
- `eval_on_start`: False
|
| 385 |
+
- `use_liger_kernel`: False
|
| 386 |
+
- `eval_use_gather_object`: False
|
| 387 |
+
- `average_tokens_across_devices`: False
|
| 388 |
+
- `prompts`: None
|
| 389 |
+
- `batch_sampler`: batch_sampler
|
| 390 |
+
- `multi_dataset_batch_sampler`: proportional
|
| 391 |
+
|
| 392 |
+
</details>
|
| 393 |
+
|
| 394 |
+
### Training Logs
|
| 395 |
+
| Epoch | Step | Training Loss | Validation Loss | sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap |
|
| 396 |
+
|:------:|:----:|:-------------:|:---------------:|:---------------------------------------------------------------------:|
|
| 397 |
+
| -1 | -1 | - | - | 0.7140 |
|
| 398 |
+
| 0.7207 | 500 | 0.038 | - | - |
|
| 399 |
+
| 0.9989 | 693 | - | 0.0028 | 0.9911 |
|
| 400 |
+
| 1.4425 | 1000 | 0.0128 | - | - |
|
| 401 |
+
| 1.9989 | 1386 | - | 0.0021 | 0.9956 |
|
| 402 |
+
| 2.1643 | 1500 | 0.0084 | - | - |
|
| 403 |
+
| 2.8850 | 2000 | 0.0065 | - | - |
|
| 404 |
+
| 2.9989 | 2079 | - | 0.0015 | 0.9968 |
|
| 405 |
+
|
| 406 |
+
|
| 407 |
+
### Framework Versions
|
| 408 |
+
- Python: 3.12.9
|
| 409 |
+
- Sentence Transformers: 3.4.1
|
| 410 |
+
- Transformers: 4.51.3
|
| 411 |
+
- PyTorch: 2.7.0+cu126
|
| 412 |
+
- Accelerate: 1.6.0
|
| 413 |
+
- Datasets: 3.6.0
|
| 414 |
+
- Tokenizers: 0.21.1
|
| 415 |
+
|
| 416 |
+
## Citation
|
| 417 |
+
|
| 418 |
+
### BibTeX
|
| 419 |
+
|
| 420 |
+
#### Sentence Transformers
|
| 421 |
+
```bibtex
|
| 422 |
+
@inproceedings{reimers-2019-sentence-bert,
|
| 423 |
+
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
|
| 424 |
+
author = "Reimers, Nils and Gurevych, Iryna",
|
| 425 |
+
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
|
| 426 |
+
month = "11",
|
| 427 |
+
year = "2019",
|
| 428 |
+
publisher = "Association for Computational Linguistics",
|
| 429 |
+
url = "https://arxiv.org/abs/1908.10084",
|
| 430 |
+
}
|
| 431 |
+
```
|
| 432 |
+
|
| 433 |
+
#### ContrastiveLoss
|
| 434 |
+
```bibtex
|
| 435 |
+
@inproceedings{hadsell2006dimensionality,
|
| 436 |
+
author={Hadsell, R. and Chopra, S. and LeCun, Y.},
|
| 437 |
+
booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)},
|
| 438 |
+
title={Dimensionality Reduction by Learning an Invariant Mapping},
|
| 439 |
+
year={2006},
|
| 440 |
+
volume={2},
|
| 441 |
+
number={},
|
| 442 |
+
pages={1735-1742},
|
| 443 |
+
doi={10.1109/CVPR.2006.100}
|
| 444 |
+
}
|
| 445 |
+
```
|
| 446 |
+
|
| 447 |
+
<!--
|
| 448 |
+
## Glossary
|
| 449 |
+
|
| 450 |
+
*Clearly define terms in order to be accessible across audiences.*
|
| 451 |
+
-->
|
| 452 |
+
|
| 453 |
+
<!--
|
| 454 |
+
## Model Card Authors
|
| 455 |
+
|
| 456 |
+
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
|
| 457 |
+
-->
|
| 458 |
+
|
| 459 |
+
<!--
|
| 460 |
+
## Model Card Contact
|
| 461 |
+
|
| 462 |
+
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
|
| 463 |
+
-->
|
checkpoint-2079/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertModel"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"gradient_checkpointing": false,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 384,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 1536,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.51.3",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 250037
|
| 25 |
+
}
|
checkpoint-2079/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.4.1",
|
| 4 |
+
"transformers": "4.51.3",
|
| 5 |
+
"pytorch": "2.7.0+cu126"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": "cosine"
|
| 10 |
+
}
|
checkpoint-2079/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:be4aae24e60c6b621f1ba97cb0d24aabcab5321c4240d46d5edb7e80659934ee
|
| 3 |
+
size 470637416
|
checkpoint-2079/modules.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
]
|
checkpoint-2079/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c9e098921ba43359be5ed80e5ba27c5533b3a15897c5d1e31312ca9a0532d345
|
| 3 |
+
size 1715019
|
checkpoint-2079/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:77bc5024a787692a4d7f38816d8d4cebfac75844ad0e708d5509da8b49d34e47
|
| 3 |
+
size 14645
|
checkpoint-2079/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f10354d1d65a8ea0cab60b6880392a6fdef91bdc8d2eafbb18ca1599d3a27f7d
|
| 3 |
+
size 1383
|
checkpoint-2079/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:4f50b78397a5b4f9a0d3308e8b1ad396b4267f9f2fa04a1619312b44f7b2ef77
|
| 3 |
+
size 1465
|
checkpoint-2079/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 128,
|
| 3 |
+
"do_lower_case": false
|
| 4 |
+
}
|
checkpoint-2079/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
checkpoint-2079/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cad551d5600a84242d0973327029452a1e3672ba6313c2a3c3d69c4310e12719
|
| 3 |
+
size 17082987
|
checkpoint-2079/tokenizer_config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"250001": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"eos_token": "</s>",
|
| 49 |
+
"extra_special_tokens": {},
|
| 50 |
+
"mask_token": "<mask>",
|
| 51 |
+
"max_length": 128,
|
| 52 |
+
"model_max_length": 128,
|
| 53 |
+
"pad_to_multiple_of": null,
|
| 54 |
+
"pad_token": "<pad>",
|
| 55 |
+
"pad_token_type_id": 0,
|
| 56 |
+
"padding_side": "right",
|
| 57 |
+
"sep_token": "</s>",
|
| 58 |
+
"stride": 0,
|
| 59 |
+
"strip_accents": null,
|
| 60 |
+
"tokenize_chinese_chars": true,
|
| 61 |
+
"tokenizer_class": "BertTokenizer",
|
| 62 |
+
"truncation_side": "right",
|
| 63 |
+
"truncation_strategy": "longest_first",
|
| 64 |
+
"unk_token": "<unk>"
|
| 65 |
+
}
|
checkpoint-2079/trainer_state.json
ADDED
|
@@ -0,0 +1,119 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2079,
|
| 3 |
+
"best_metric": 0.0014588695485144854,
|
| 4 |
+
"best_model_checkpoint": "data/fine-tuned-sbert-sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2-original-adafactor/checkpoint-2079",
|
| 5 |
+
"epoch": 2.998918918918919,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 2079,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.7207207207207207,
|
| 14 |
+
"grad_norm": 0.15965162217617035,
|
| 15 |
+
"learning_rate": 2.7341619887730554e-05,
|
| 16 |
+
"loss": 0.038,
|
| 17 |
+
"step": 500
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.9989189189189189,
|
| 21 |
+
"eval_loss": 0.002777691464871168,
|
| 22 |
+
"eval_runtime": 792.7494,
|
| 23 |
+
"eval_samples_per_second": 3359.543,
|
| 24 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.975327415818089,
|
| 25 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.7701693773269653,
|
| 26 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.9911172257655966,
|
| 27 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9624723810391563,
|
| 28 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.7597355842590332,
|
| 29 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9441287064717891,
|
| 30 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9527175567355498,
|
| 31 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.9724290300680068,
|
| 32 |
+
"eval_steps_per_second": 4.375,
|
| 33 |
+
"step": 693
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 1.4425225225225224,
|
| 37 |
+
"grad_norm": 0.11966603249311447,
|
| 38 |
+
"learning_rate": 2.1327185244587008e-05,
|
| 39 |
+
"loss": 0.0128,
|
| 40 |
+
"step": 1000
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 1.998918918918919,
|
| 44 |
+
"eval_loss": 0.002118302509188652,
|
| 45 |
+
"eval_runtime": 789.0568,
|
| 46 |
+
"eval_samples_per_second": 3375.265,
|
| 47 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.9846242227629088,
|
| 48 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.6801187992095947,
|
| 49 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.9955904030209028,
|
| 50 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9765449140552956,
|
| 51 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.6780189275741577,
|
| 52 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9651303277408154,
|
| 53 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9721848413657824,
|
| 54 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.9809442711989229,
|
| 55 |
+
"eval_steps_per_second": 4.395,
|
| 56 |
+
"step": 1386
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.1643243243243244,
|
| 60 |
+
"grad_norm": 0.04637068510055542,
|
| 61 |
+
"learning_rate": 1.5312750601443466e-05,
|
| 62 |
+
"loss": 0.0084,
|
| 63 |
+
"step": 1500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.885045045045045,
|
| 67 |
+
"grad_norm": 0.07652924209833145,
|
| 68 |
+
"learning_rate": 9.29831595829992e-06,
|
| 69 |
+
"loss": 0.0065,
|
| 70 |
+
"step": 2000
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.998918918918919,
|
| 74 |
+
"eval_loss": 0.0014588695485144854,
|
| 75 |
+
"eval_runtime": 782.4422,
|
| 76 |
+
"eval_samples_per_second": 3403.799,
|
| 77 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.9879171547865789,
|
| 78 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.7181636691093445,
|
| 79 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.996840725826042,
|
| 80 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9815604299892273,
|
| 81 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.7181636691093445,
|
| 82 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9725931427811844,
|
| 83 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9775832353646149,
|
| 84 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.98557011840788,
|
| 85 |
+
"eval_steps_per_second": 4.432,
|
| 86 |
+
"step": 2079
|
| 87 |
+
}
|
| 88 |
+
],
|
| 89 |
+
"logging_steps": 500,
|
| 90 |
+
"max_steps": 2772,
|
| 91 |
+
"num_input_tokens_seen": 0,
|
| 92 |
+
"num_train_epochs": 4,
|
| 93 |
+
"save_steps": 100,
|
| 94 |
+
"stateful_callbacks": {
|
| 95 |
+
"EarlyStoppingCallback": {
|
| 96 |
+
"args": {
|
| 97 |
+
"early_stopping_patience": 1,
|
| 98 |
+
"early_stopping_threshold": 0.0
|
| 99 |
+
},
|
| 100 |
+
"attributes": {
|
| 101 |
+
"early_stopping_patience_counter": 0
|
| 102 |
+
}
|
| 103 |
+
},
|
| 104 |
+
"TrainerControl": {
|
| 105 |
+
"args": {
|
| 106 |
+
"should_epoch_stop": false,
|
| 107 |
+
"should_evaluate": false,
|
| 108 |
+
"should_log": false,
|
| 109 |
+
"should_save": true,
|
| 110 |
+
"should_training_stop": false
|
| 111 |
+
},
|
| 112 |
+
"attributes": {}
|
| 113 |
+
}
|
| 114 |
+
},
|
| 115 |
+
"total_flos": 0.0,
|
| 116 |
+
"train_batch_size": 768,
|
| 117 |
+
"trial_name": null,
|
| 118 |
+
"trial_params": null
|
| 119 |
+
}
|
checkpoint-2079/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee490aac80277022fdde76dd6763dbabee88802149f3f246e7d0572b504fcdaf
|
| 3 |
+
size 6097
|
checkpoint-2079/unigram.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:da145b5e7700ae40f16691ec32a0b1fdc1ee3298db22a31ea55f57a966c4a65d
|
| 3 |
+
size 14763260
|
checkpoint-2772/1_Pooling/config.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"word_embedding_dimension": 384,
|
| 3 |
+
"pooling_mode_cls_token": false,
|
| 4 |
+
"pooling_mode_mean_tokens": true,
|
| 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 |
+
}
|
checkpoint-2772/README.md
ADDED
|
@@ -0,0 +1,465 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- en
|
| 4 |
+
license: apache-2.0
|
| 5 |
+
tags:
|
| 6 |
+
- sentence-transformers
|
| 7 |
+
- sentence-similarity
|
| 8 |
+
- feature-extraction
|
| 9 |
+
- generated_from_trainer
|
| 10 |
+
- dataset_size:2130621
|
| 11 |
+
- loss:ContrastiveLoss
|
| 12 |
+
base_model: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
|
| 13 |
+
widget:
|
| 14 |
+
- source_sentence: Kim Chol-sam
|
| 15 |
+
sentences:
|
| 16 |
+
- Stankevich Sergey Nikolayevich
|
| 17 |
+
- Kim Chin-So’k
|
| 18 |
+
- Julen Lopetegui Agote
|
| 19 |
+
- source_sentence: دينا بنت عبد الحميد
|
| 20 |
+
sentences:
|
| 21 |
+
- Alexia van Amsberg
|
| 22 |
+
- Anthony Nicholas Colin Maitland Biddulph, 5th Baron Biddulph
|
| 23 |
+
- Dina bint Abdul-Hamíd
|
| 24 |
+
- source_sentence: Մուհամեդ բեն Նաիֆ Ալ Սաուդ
|
| 25 |
+
sentences:
|
| 26 |
+
- Karpov Anatoly Evgenyevich
|
| 27 |
+
- GNPower Mariveles Coal Plant [former]
|
| 28 |
+
- Muhammed bin Nayef bin Abdul Aziz Al Saud
|
| 29 |
+
- source_sentence: Edward Gnehm
|
| 30 |
+
sentences:
|
| 31 |
+
- Шауэрте, Хартмут
|
| 32 |
+
- Ханзада Филипп, Эдинбург герцогі
|
| 33 |
+
- AFX
|
| 34 |
+
- source_sentence: Schori i Lidingö
|
| 35 |
+
sentences:
|
| 36 |
+
- Yordan Canev
|
| 37 |
+
- ကားပေါ့ အန်နာတိုလီ
|
| 38 |
+
- BYSTROV, Mikhail Ivanovich
|
| 39 |
+
pipeline_tag: sentence-similarity
|
| 40 |
+
library_name: sentence-transformers
|
| 41 |
+
metrics:
|
| 42 |
+
- cosine_accuracy
|
| 43 |
+
- cosine_accuracy_threshold
|
| 44 |
+
- cosine_f1
|
| 45 |
+
- cosine_f1_threshold
|
| 46 |
+
- cosine_precision
|
| 47 |
+
- cosine_recall
|
| 48 |
+
- cosine_ap
|
| 49 |
+
- cosine_mcc
|
| 50 |
+
model-index:
|
| 51 |
+
- name: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2-name-matcher-original
|
| 52 |
+
results:
|
| 53 |
+
- task:
|
| 54 |
+
type: binary-classification
|
| 55 |
+
name: Binary Classification
|
| 56 |
+
dataset:
|
| 57 |
+
name: sentence transformers paraphrase multilingual MiniLM L12 v2
|
| 58 |
+
type: sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2
|
| 59 |
+
metrics:
|
| 60 |
+
- type: cosine_accuracy
|
| 61 |
+
value: 0.9885216725241056
|
| 62 |
+
name: Cosine Accuracy
|
| 63 |
+
- type: cosine_accuracy_threshold
|
| 64 |
+
value: 0.7183246612548828
|
| 65 |
+
name: Cosine Accuracy Threshold
|
| 66 |
+
- type: cosine_f1
|
| 67 |
+
value: 0.9824706124974221
|
| 68 |
+
name: Cosine F1
|
| 69 |
+
- type: cosine_f1_threshold
|
| 70 |
+
value: 0.7085607051849365
|
| 71 |
+
name: Cosine F1 Threshold
|
| 72 |
+
- type: cosine_precision
|
| 73 |
+
value: 0.9782229269572558
|
| 74 |
+
name: Cosine Precision
|
| 75 |
+
- type: cosine_recall
|
| 76 |
+
value: 0.9867553479166427
|
| 77 |
+
name: Cosine Recall
|
| 78 |
+
- type: cosine_ap
|
| 79 |
+
value: 0.9971022799526896
|
| 80 |
+
name: Cosine Ap
|
| 81 |
+
- type: cosine_mcc
|
| 82 |
+
value: 0.9739458779668466
|
| 83 |
+
name: Cosine Mcc
|
| 84 |
+
---
|
| 85 |
+
|
| 86 |
+
# sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2-name-matcher-original
|
| 87 |
+
|
| 88 |
+
This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2). It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.
|
| 89 |
+
|
| 90 |
+
## Model Details
|
| 91 |
+
|
| 92 |
+
### Model Description
|
| 93 |
+
- **Model Type:** Sentence Transformer
|
| 94 |
+
- **Base model:** [sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2](https://huggingface.co/sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2) <!-- at revision 86741b4e3f5cb7765a600d3a3d55a0f6a6cb443d -->
|
| 95 |
+
- **Maximum Sequence Length:** 128 tokens
|
| 96 |
+
- **Output Dimensionality:** 384 dimensions
|
| 97 |
+
- **Similarity Function:** Cosine Similarity
|
| 98 |
+
<!-- - **Training Dataset:** Unknown -->
|
| 99 |
+
- **Language:** en
|
| 100 |
+
- **License:** apache-2.0
|
| 101 |
+
|
| 102 |
+
### Model Sources
|
| 103 |
+
|
| 104 |
+
- **Documentation:** [Sentence Transformers Documentation](https://sbert.net)
|
| 105 |
+
- **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/sentence-transformers)
|
| 106 |
+
- **Hugging Face:** [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)
|
| 107 |
+
|
| 108 |
+
### Full Model Architecture
|
| 109 |
+
|
| 110 |
+
```
|
| 111 |
+
SentenceTransformer(
|
| 112 |
+
(0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel
|
| 113 |
+
(1): Pooling({'word_embedding_dimension': 384, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
|
| 114 |
+
)
|
| 115 |
+
```
|
| 116 |
+
|
| 117 |
+
## Usage
|
| 118 |
+
|
| 119 |
+
### Direct Usage (Sentence Transformers)
|
| 120 |
+
|
| 121 |
+
First install the Sentence Transformers library:
|
| 122 |
+
|
| 123 |
+
```bash
|
| 124 |
+
pip install -U sentence-transformers
|
| 125 |
+
```
|
| 126 |
+
|
| 127 |
+
Then you can load this model and run inference.
|
| 128 |
+
```python
|
| 129 |
+
from sentence_transformers import SentenceTransformer
|
| 130 |
+
|
| 131 |
+
# Download from the 🤗 Hub
|
| 132 |
+
model = SentenceTransformer("sentence_transformers_model_id")
|
| 133 |
+
# Run inference
|
| 134 |
+
sentences = [
|
| 135 |
+
'Schori i Lidingö',
|
| 136 |
+
'Yordan Canev',
|
| 137 |
+
'ကားပေါ့ အန်နာတိုလီ',
|
| 138 |
+
]
|
| 139 |
+
embeddings = model.encode(sentences)
|
| 140 |
+
print(embeddings.shape)
|
| 141 |
+
# [3, 384]
|
| 142 |
+
|
| 143 |
+
# Get the similarity scores for the embeddings
|
| 144 |
+
similarities = model.similarity(embeddings, embeddings)
|
| 145 |
+
print(similarities.shape)
|
| 146 |
+
# [3, 3]
|
| 147 |
+
```
|
| 148 |
+
|
| 149 |
+
<!--
|
| 150 |
+
### Direct Usage (Transformers)
|
| 151 |
+
|
| 152 |
+
<details><summary>Click to see the direct usage in Transformers</summary>
|
| 153 |
+
|
| 154 |
+
</details>
|
| 155 |
+
-->
|
| 156 |
+
|
| 157 |
+
<!--
|
| 158 |
+
### Downstream Usage (Sentence Transformers)
|
| 159 |
+
|
| 160 |
+
You can finetune this model on your own dataset.
|
| 161 |
+
|
| 162 |
+
<details><summary>Click to expand</summary>
|
| 163 |
+
|
| 164 |
+
</details>
|
| 165 |
+
-->
|
| 166 |
+
|
| 167 |
+
<!--
|
| 168 |
+
### Out-of-Scope Use
|
| 169 |
+
|
| 170 |
+
*List how the model may foreseeably be misused and address what users ought not to do with the model.*
|
| 171 |
+
-->
|
| 172 |
+
|
| 173 |
+
## Evaluation
|
| 174 |
+
|
| 175 |
+
### Metrics
|
| 176 |
+
|
| 177 |
+
#### Binary Classification
|
| 178 |
+
|
| 179 |
+
* Dataset: `sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2`
|
| 180 |
+
* Evaluated with [<code>BinaryClassificationEvaluator</code>](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.BinaryClassificationEvaluator)
|
| 181 |
+
|
| 182 |
+
| Metric | Value |
|
| 183 |
+
|:--------------------------|:-----------|
|
| 184 |
+
| cosine_accuracy | 0.9885 |
|
| 185 |
+
| cosine_accuracy_threshold | 0.7183 |
|
| 186 |
+
| cosine_f1 | 0.9825 |
|
| 187 |
+
| cosine_f1_threshold | 0.7086 |
|
| 188 |
+
| cosine_precision | 0.9782 |
|
| 189 |
+
| cosine_recall | 0.9868 |
|
| 190 |
+
| **cosine_ap** | **0.9971** |
|
| 191 |
+
| cosine_mcc | 0.9739 |
|
| 192 |
+
|
| 193 |
+
<!--
|
| 194 |
+
## Bias, Risks and Limitations
|
| 195 |
+
|
| 196 |
+
*What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.*
|
| 197 |
+
-->
|
| 198 |
+
|
| 199 |
+
<!--
|
| 200 |
+
### Recommendations
|
| 201 |
+
|
| 202 |
+
*What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.*
|
| 203 |
+
-->
|
| 204 |
+
|
| 205 |
+
## Training Details
|
| 206 |
+
|
| 207 |
+
### Training Dataset
|
| 208 |
+
|
| 209 |
+
#### Unnamed Dataset
|
| 210 |
+
|
| 211 |
+
* Size: 2,130,621 training samples
|
| 212 |
+
* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
|
| 213 |
+
* Approximate statistics based on the first 1000 samples:
|
| 214 |
+
| | sentence1 | sentence2 | label |
|
| 215 |
+
|:--------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------------------------|:---------------------------------------------------------------|
|
| 216 |
+
| type | string | string | float |
|
| 217 |
+
| details | <ul><li>min: 3 tokens</li><li>mean: 9.32 tokens</li><li>max: 57 tokens</li></ul> | <ul><li>min: 3 tokens</li><li>mean: 9.16 tokens</li><li>max: 54 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.34</li><li>max: 1.0</li></ul> |
|
| 218 |
+
* Samples:
|
| 219 |
+
| sentence1 | sentence2 | label |
|
| 220 |
+
|:----------------------------------|:------------------------------------|:-----------------|
|
| 221 |
+
| <code>캐스린 설리번</code> | <code>Kathryn D. Sullivanová</code> | <code>1.0</code> |
|
| 222 |
+
| <code>ଶିବରାଜ ଅଧାଲରାଓ ପାଟିଲ</code> | <code>Aleksander Lubocki</code> | <code>0.0</code> |
|
| 223 |
+
| <code>Пырванов, Георги</code> | <code>アナトーリー・セルジュコフ</code> | <code>0.0</code> |
|
| 224 |
+
* Loss: [<code>ContrastiveLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastiveloss) with these parameters:
|
| 225 |
+
```json
|
| 226 |
+
{
|
| 227 |
+
"distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
|
| 228 |
+
"margin": 0.5,
|
| 229 |
+
"size_average": true
|
| 230 |
+
}
|
| 231 |
+
```
|
| 232 |
+
|
| 233 |
+
### Evaluation Dataset
|
| 234 |
+
|
| 235 |
+
#### Unnamed Dataset
|
| 236 |
+
|
| 237 |
+
* Size: 2,663,276 evaluation samples
|
| 238 |
+
* Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
|
| 239 |
+
* Approximate statistics based on the first 1000 samples:
|
| 240 |
+
| | sentence1 | sentence2 | label |
|
| 241 |
+
|:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:---------------------------------------------------------------|
|
| 242 |
+
| type | string | string | float |
|
| 243 |
+
| details | <ul><li>min: 3 tokens</li><li>mean: 9.34 tokens</li><li>max: 102 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 9.11 tokens</li><li>max: 100 tokens</li></ul> | <ul><li>min: 0.0</li><li>mean: 0.33</li><li>max: 1.0</li></ul> |
|
| 244 |
+
* Samples:
|
| 245 |
+
| sentence1 | sentence2 | label |
|
| 246 |
+
|:--------------------------------------|:---------------------------------------|:-----------------|
|
| 247 |
+
| <code>Ева Херман</code> | <code>I Xuan Karlos</code> | <code>0.0</code> |
|
| 248 |
+
| <code>Кличков Андрій Євгенович</code> | <code>Андрэй Яўгенавіч Клычкоў</code> | <code>1.0</code> |
|
| 249 |
+
| <code>Кинах А.</code> | <code>Senator John Hickenlooper</code> | <code>0.0</code> |
|
| 250 |
+
* Loss: [<code>ContrastiveLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#contrastiveloss) with these parameters:
|
| 251 |
+
```json
|
| 252 |
+
{
|
| 253 |
+
"distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
|
| 254 |
+
"margin": 0.5,
|
| 255 |
+
"size_average": true
|
| 256 |
+
}
|
| 257 |
+
```
|
| 258 |
+
|
| 259 |
+
### Training Hyperparameters
|
| 260 |
+
#### Non-Default Hyperparameters
|
| 261 |
+
|
| 262 |
+
- `eval_strategy`: epoch
|
| 263 |
+
- `per_device_train_batch_size`: 768
|
| 264 |
+
- `per_device_eval_batch_size`: 768
|
| 265 |
+
- `gradient_accumulation_steps`: 4
|
| 266 |
+
- `learning_rate`: 3e-05
|
| 267 |
+
- `weight_decay`: 0.01
|
| 268 |
+
- `num_train_epochs`: 4
|
| 269 |
+
- `warmup_ratio`: 0.1
|
| 270 |
+
- `fp16`: True
|
| 271 |
+
- `load_best_model_at_end`: True
|
| 272 |
+
- `optim`: adafactor
|
| 273 |
+
|
| 274 |
+
#### All Hyperparameters
|
| 275 |
+
<details><summary>Click to expand</summary>
|
| 276 |
+
|
| 277 |
+
- `overwrite_output_dir`: False
|
| 278 |
+
- `do_predict`: False
|
| 279 |
+
- `eval_strategy`: epoch
|
| 280 |
+
- `prediction_loss_only`: True
|
| 281 |
+
- `per_device_train_batch_size`: 768
|
| 282 |
+
- `per_device_eval_batch_size`: 768
|
| 283 |
+
- `per_gpu_train_batch_size`: None
|
| 284 |
+
- `per_gpu_eval_batch_size`: None
|
| 285 |
+
- `gradient_accumulation_steps`: 4
|
| 286 |
+
- `eval_accumulation_steps`: None
|
| 287 |
+
- `torch_empty_cache_steps`: None
|
| 288 |
+
- `learning_rate`: 3e-05
|
| 289 |
+
- `weight_decay`: 0.01
|
| 290 |
+
- `adam_beta1`: 0.9
|
| 291 |
+
- `adam_beta2`: 0.999
|
| 292 |
+
- `adam_epsilon`: 1e-08
|
| 293 |
+
- `max_grad_norm`: 1.0
|
| 294 |
+
- `num_train_epochs`: 4
|
| 295 |
+
- `max_steps`: -1
|
| 296 |
+
- `lr_scheduler_type`: linear
|
| 297 |
+
- `lr_scheduler_kwargs`: {}
|
| 298 |
+
- `warmup_ratio`: 0.1
|
| 299 |
+
- `warmup_steps`: 0
|
| 300 |
+
- `log_level`: passive
|
| 301 |
+
- `log_level_replica`: warning
|
| 302 |
+
- `log_on_each_node`: True
|
| 303 |
+
- `logging_nan_inf_filter`: True
|
| 304 |
+
- `save_safetensors`: True
|
| 305 |
+
- `save_on_each_node`: False
|
| 306 |
+
- `save_only_model`: False
|
| 307 |
+
- `restore_callback_states_from_checkpoint`: False
|
| 308 |
+
- `no_cuda`: False
|
| 309 |
+
- `use_cpu`: False
|
| 310 |
+
- `use_mps_device`: False
|
| 311 |
+
- `seed`: 42
|
| 312 |
+
- `data_seed`: None
|
| 313 |
+
- `jit_mode_eval`: False
|
| 314 |
+
- `use_ipex`: False
|
| 315 |
+
- `bf16`: False
|
| 316 |
+
- `fp16`: True
|
| 317 |
+
- `fp16_opt_level`: O1
|
| 318 |
+
- `half_precision_backend`: auto
|
| 319 |
+
- `bf16_full_eval`: False
|
| 320 |
+
- `fp16_full_eval`: False
|
| 321 |
+
- `tf32`: None
|
| 322 |
+
- `local_rank`: 0
|
| 323 |
+
- `ddp_backend`: None
|
| 324 |
+
- `tpu_num_cores`: None
|
| 325 |
+
- `tpu_metrics_debug`: False
|
| 326 |
+
- `debug`: []
|
| 327 |
+
- `dataloader_drop_last`: False
|
| 328 |
+
- `dataloader_num_workers`: 0
|
| 329 |
+
- `dataloader_prefetch_factor`: None
|
| 330 |
+
- `past_index`: -1
|
| 331 |
+
- `disable_tqdm`: False
|
| 332 |
+
- `remove_unused_columns`: True
|
| 333 |
+
- `label_names`: None
|
| 334 |
+
- `load_best_model_at_end`: True
|
| 335 |
+
- `ignore_data_skip`: False
|
| 336 |
+
- `fsdp`: []
|
| 337 |
+
- `fsdp_min_num_params`: 0
|
| 338 |
+
- `fsdp_config`: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
|
| 339 |
+
- `tp_size`: 0
|
| 340 |
+
- `fsdp_transformer_layer_cls_to_wrap`: None
|
| 341 |
+
- `accelerator_config`: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
|
| 342 |
+
- `deepspeed`: None
|
| 343 |
+
- `label_smoothing_factor`: 0.0
|
| 344 |
+
- `optim`: adafactor
|
| 345 |
+
- `optim_args`: None
|
| 346 |
+
- `adafactor`: False
|
| 347 |
+
- `group_by_length`: False
|
| 348 |
+
- `length_column_name`: length
|
| 349 |
+
- `ddp_find_unused_parameters`: None
|
| 350 |
+
- `ddp_bucket_cap_mb`: None
|
| 351 |
+
- `ddp_broadcast_buffers`: False
|
| 352 |
+
- `dataloader_pin_memory`: True
|
| 353 |
+
- `dataloader_persistent_workers`: False
|
| 354 |
+
- `skip_memory_metrics`: True
|
| 355 |
+
- `use_legacy_prediction_loop`: False
|
| 356 |
+
- `push_to_hub`: False
|
| 357 |
+
- `resume_from_checkpoint`: None
|
| 358 |
+
- `hub_model_id`: None
|
| 359 |
+
- `hub_strategy`: every_save
|
| 360 |
+
- `hub_private_repo`: None
|
| 361 |
+
- `hub_always_push`: False
|
| 362 |
+
- `gradient_checkpointing`: False
|
| 363 |
+
- `gradient_checkpointing_kwargs`: None
|
| 364 |
+
- `include_inputs_for_metrics`: False
|
| 365 |
+
- `include_for_metrics`: []
|
| 366 |
+
- `eval_do_concat_batches`: True
|
| 367 |
+
- `fp16_backend`: auto
|
| 368 |
+
- `push_to_hub_model_id`: None
|
| 369 |
+
- `push_to_hub_organization`: None
|
| 370 |
+
- `mp_parameters`:
|
| 371 |
+
- `auto_find_batch_size`: False
|
| 372 |
+
- `full_determinism`: False
|
| 373 |
+
- `torchdynamo`: None
|
| 374 |
+
- `ray_scope`: last
|
| 375 |
+
- `ddp_timeout`: 1800
|
| 376 |
+
- `torch_compile`: False
|
| 377 |
+
- `torch_compile_backend`: None
|
| 378 |
+
- `torch_compile_mode`: None
|
| 379 |
+
- `include_tokens_per_second`: False
|
| 380 |
+
- `include_num_input_tokens_seen`: False
|
| 381 |
+
- `neftune_noise_alpha`: None
|
| 382 |
+
- `optim_target_modules`: None
|
| 383 |
+
- `batch_eval_metrics`: False
|
| 384 |
+
- `eval_on_start`: False
|
| 385 |
+
- `use_liger_kernel`: False
|
| 386 |
+
- `eval_use_gather_object`: False
|
| 387 |
+
- `average_tokens_across_devices`: False
|
| 388 |
+
- `prompts`: None
|
| 389 |
+
- `batch_sampler`: batch_sampler
|
| 390 |
+
- `multi_dataset_batch_sampler`: proportional
|
| 391 |
+
|
| 392 |
+
</details>
|
| 393 |
+
|
| 394 |
+
### Training Logs
|
| 395 |
+
| Epoch | Step | Training Loss | Validation Loss | sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap |
|
| 396 |
+
|:------:|:----:|:-------------:|:---------------:|:---------------------------------------------------------------------:|
|
| 397 |
+
| -1 | -1 | - | - | 0.7140 |
|
| 398 |
+
| 0.7207 | 500 | 0.038 | - | - |
|
| 399 |
+
| 0.9989 | 693 | - | 0.0028 | 0.9911 |
|
| 400 |
+
| 1.4425 | 1000 | 0.0128 | - | - |
|
| 401 |
+
| 1.9989 | 1386 | - | 0.0021 | 0.9956 |
|
| 402 |
+
| 2.1643 | 1500 | 0.0084 | - | - |
|
| 403 |
+
| 2.8850 | 2000 | 0.0065 | - | - |
|
| 404 |
+
| 2.9989 | 2079 | - | 0.0015 | 0.9968 |
|
| 405 |
+
| 3.6068 | 2500 | 0.0056 | - | - |
|
| 406 |
+
| 3.9989 | 2772 | - | 0.0014 | 0.9971 |
|
| 407 |
+
|
| 408 |
+
|
| 409 |
+
### Framework Versions
|
| 410 |
+
- Python: 3.12.9
|
| 411 |
+
- Sentence Transformers: 3.4.1
|
| 412 |
+
- Transformers: 4.51.3
|
| 413 |
+
- PyTorch: 2.7.0+cu126
|
| 414 |
+
- Accelerate: 1.6.0
|
| 415 |
+
- Datasets: 3.6.0
|
| 416 |
+
- Tokenizers: 0.21.1
|
| 417 |
+
|
| 418 |
+
## Citation
|
| 419 |
+
|
| 420 |
+
### BibTeX
|
| 421 |
+
|
| 422 |
+
#### Sentence Transformers
|
| 423 |
+
```bibtex
|
| 424 |
+
@inproceedings{reimers-2019-sentence-bert,
|
| 425 |
+
title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
|
| 426 |
+
author = "Reimers, Nils and Gurevych, Iryna",
|
| 427 |
+
booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
|
| 428 |
+
month = "11",
|
| 429 |
+
year = "2019",
|
| 430 |
+
publisher = "Association for Computational Linguistics",
|
| 431 |
+
url = "https://arxiv.org/abs/1908.10084",
|
| 432 |
+
}
|
| 433 |
+
```
|
| 434 |
+
|
| 435 |
+
#### ContrastiveLoss
|
| 436 |
+
```bibtex
|
| 437 |
+
@inproceedings{hadsell2006dimensionality,
|
| 438 |
+
author={Hadsell, R. and Chopra, S. and LeCun, Y.},
|
| 439 |
+
booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)},
|
| 440 |
+
title={Dimensionality Reduction by Learning an Invariant Mapping},
|
| 441 |
+
year={2006},
|
| 442 |
+
volume={2},
|
| 443 |
+
number={},
|
| 444 |
+
pages={1735-1742},
|
| 445 |
+
doi={10.1109/CVPR.2006.100}
|
| 446 |
+
}
|
| 447 |
+
```
|
| 448 |
+
|
| 449 |
+
<!--
|
| 450 |
+
## Glossary
|
| 451 |
+
|
| 452 |
+
*Clearly define terms in order to be accessible across audiences.*
|
| 453 |
+
-->
|
| 454 |
+
|
| 455 |
+
<!--
|
| 456 |
+
## Model Card Authors
|
| 457 |
+
|
| 458 |
+
*Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.*
|
| 459 |
+
-->
|
| 460 |
+
|
| 461 |
+
<!--
|
| 462 |
+
## Model Card Contact
|
| 463 |
+
|
| 464 |
+
*Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.*
|
| 465 |
+
-->
|
checkpoint-2772/config.json
ADDED
|
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"BertModel"
|
| 4 |
+
],
|
| 5 |
+
"attention_probs_dropout_prob": 0.1,
|
| 6 |
+
"classifier_dropout": null,
|
| 7 |
+
"gradient_checkpointing": false,
|
| 8 |
+
"hidden_act": "gelu",
|
| 9 |
+
"hidden_dropout_prob": 0.1,
|
| 10 |
+
"hidden_size": 384,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 1536,
|
| 13 |
+
"layer_norm_eps": 1e-12,
|
| 14 |
+
"max_position_embeddings": 512,
|
| 15 |
+
"model_type": "bert",
|
| 16 |
+
"num_attention_heads": 12,
|
| 17 |
+
"num_hidden_layers": 12,
|
| 18 |
+
"pad_token_id": 0,
|
| 19 |
+
"position_embedding_type": "absolute",
|
| 20 |
+
"torch_dtype": "float32",
|
| 21 |
+
"transformers_version": "4.51.3",
|
| 22 |
+
"type_vocab_size": 2,
|
| 23 |
+
"use_cache": true,
|
| 24 |
+
"vocab_size": 250037
|
| 25 |
+
}
|
checkpoint-2772/config_sentence_transformers.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"__version__": {
|
| 3 |
+
"sentence_transformers": "3.4.1",
|
| 4 |
+
"transformers": "4.51.3",
|
| 5 |
+
"pytorch": "2.7.0+cu126"
|
| 6 |
+
},
|
| 7 |
+
"prompts": {},
|
| 8 |
+
"default_prompt_name": null,
|
| 9 |
+
"similarity_fn_name": "cosine"
|
| 10 |
+
}
|
checkpoint-2772/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:86cbcc5809568045ff4d80ea5eff6ddee83cd424541447967273a7023c7e187e
|
| 3 |
+
size 470637416
|
checkpoint-2772/modules.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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 |
+
]
|
checkpoint-2772/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ac3e5f9bb09ee4343f28e1fd561f33a38c1f23ae8dae410e77420c9278ae6ac8
|
| 3 |
+
size 1715019
|
checkpoint-2772/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cd92514717bc3cb44100d9bb038875d49cd30e2d1925b33a6ddd1f8bbba8e940
|
| 3 |
+
size 14645
|
checkpoint-2772/scaler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:01d00f48ebab6e58e9ae555bb7a1dc0eb436c3b5115601ee219420aae460b550
|
| 3 |
+
size 1383
|
checkpoint-2772/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0c76e070cc219bc8a3b3a9bbab555445f78c28c96a21cc55f454628bf0d5ef07
|
| 3 |
+
size 1465
|
checkpoint-2772/sentence_bert_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"max_seq_length": 128,
|
| 3 |
+
"do_lower_case": false
|
| 4 |
+
}
|
checkpoint-2772/special_tokens_map.json
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token": {
|
| 3 |
+
"content": "<s>",
|
| 4 |
+
"lstrip": false,
|
| 5 |
+
"normalized": false,
|
| 6 |
+
"rstrip": false,
|
| 7 |
+
"single_word": false
|
| 8 |
+
},
|
| 9 |
+
"cls_token": {
|
| 10 |
+
"content": "<s>",
|
| 11 |
+
"lstrip": false,
|
| 12 |
+
"normalized": false,
|
| 13 |
+
"rstrip": false,
|
| 14 |
+
"single_word": false
|
| 15 |
+
},
|
| 16 |
+
"eos_token": {
|
| 17 |
+
"content": "</s>",
|
| 18 |
+
"lstrip": false,
|
| 19 |
+
"normalized": false,
|
| 20 |
+
"rstrip": false,
|
| 21 |
+
"single_word": false
|
| 22 |
+
},
|
| 23 |
+
"mask_token": {
|
| 24 |
+
"content": "<mask>",
|
| 25 |
+
"lstrip": true,
|
| 26 |
+
"normalized": false,
|
| 27 |
+
"rstrip": false,
|
| 28 |
+
"single_word": false
|
| 29 |
+
},
|
| 30 |
+
"pad_token": {
|
| 31 |
+
"content": "<pad>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false
|
| 36 |
+
},
|
| 37 |
+
"sep_token": {
|
| 38 |
+
"content": "</s>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false
|
| 43 |
+
},
|
| 44 |
+
"unk_token": {
|
| 45 |
+
"content": "<unk>",
|
| 46 |
+
"lstrip": false,
|
| 47 |
+
"normalized": false,
|
| 48 |
+
"rstrip": false,
|
| 49 |
+
"single_word": false
|
| 50 |
+
}
|
| 51 |
+
}
|
checkpoint-2772/tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:cad551d5600a84242d0973327029452a1e3672ba6313c2a3c3d69c4310e12719
|
| 3 |
+
size 17082987
|
checkpoint-2772/tokenizer_config.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"added_tokens_decoder": {
|
| 3 |
+
"0": {
|
| 4 |
+
"content": "<s>",
|
| 5 |
+
"lstrip": false,
|
| 6 |
+
"normalized": false,
|
| 7 |
+
"rstrip": false,
|
| 8 |
+
"single_word": false,
|
| 9 |
+
"special": true
|
| 10 |
+
},
|
| 11 |
+
"1": {
|
| 12 |
+
"content": "<pad>",
|
| 13 |
+
"lstrip": false,
|
| 14 |
+
"normalized": false,
|
| 15 |
+
"rstrip": false,
|
| 16 |
+
"single_word": false,
|
| 17 |
+
"special": true
|
| 18 |
+
},
|
| 19 |
+
"2": {
|
| 20 |
+
"content": "</s>",
|
| 21 |
+
"lstrip": false,
|
| 22 |
+
"normalized": false,
|
| 23 |
+
"rstrip": false,
|
| 24 |
+
"single_word": false,
|
| 25 |
+
"special": true
|
| 26 |
+
},
|
| 27 |
+
"3": {
|
| 28 |
+
"content": "<unk>",
|
| 29 |
+
"lstrip": false,
|
| 30 |
+
"normalized": false,
|
| 31 |
+
"rstrip": false,
|
| 32 |
+
"single_word": false,
|
| 33 |
+
"special": true
|
| 34 |
+
},
|
| 35 |
+
"250001": {
|
| 36 |
+
"content": "<mask>",
|
| 37 |
+
"lstrip": true,
|
| 38 |
+
"normalized": false,
|
| 39 |
+
"rstrip": false,
|
| 40 |
+
"single_word": false,
|
| 41 |
+
"special": true
|
| 42 |
+
}
|
| 43 |
+
},
|
| 44 |
+
"bos_token": "<s>",
|
| 45 |
+
"clean_up_tokenization_spaces": false,
|
| 46 |
+
"cls_token": "<s>",
|
| 47 |
+
"do_lower_case": true,
|
| 48 |
+
"eos_token": "</s>",
|
| 49 |
+
"extra_special_tokens": {},
|
| 50 |
+
"mask_token": "<mask>",
|
| 51 |
+
"max_length": 128,
|
| 52 |
+
"model_max_length": 128,
|
| 53 |
+
"pad_to_multiple_of": null,
|
| 54 |
+
"pad_token": "<pad>",
|
| 55 |
+
"pad_token_type_id": 0,
|
| 56 |
+
"padding_side": "right",
|
| 57 |
+
"sep_token": "</s>",
|
| 58 |
+
"stride": 0,
|
| 59 |
+
"strip_accents": null,
|
| 60 |
+
"tokenize_chinese_chars": true,
|
| 61 |
+
"tokenizer_class": "BertTokenizer",
|
| 62 |
+
"truncation_side": "right",
|
| 63 |
+
"truncation_strategy": "longest_first",
|
| 64 |
+
"unk_token": "<unk>"
|
| 65 |
+
}
|
checkpoint-2772/trainer_state.json
ADDED
|
@@ -0,0 +1,142 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_global_step": 2772,
|
| 3 |
+
"best_metric": 0.001379696768708527,
|
| 4 |
+
"best_model_checkpoint": "data/fine-tuned-sbert-sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2-original-adafactor/checkpoint-2772",
|
| 5 |
+
"epoch": 3.998918918918919,
|
| 6 |
+
"eval_steps": 100,
|
| 7 |
+
"global_step": 2772,
|
| 8 |
+
"is_hyper_param_search": false,
|
| 9 |
+
"is_local_process_zero": true,
|
| 10 |
+
"is_world_process_zero": true,
|
| 11 |
+
"log_history": [
|
| 12 |
+
{
|
| 13 |
+
"epoch": 0.7207207207207207,
|
| 14 |
+
"grad_norm": 0.15965162217617035,
|
| 15 |
+
"learning_rate": 2.7341619887730554e-05,
|
| 16 |
+
"loss": 0.038,
|
| 17 |
+
"step": 500
|
| 18 |
+
},
|
| 19 |
+
{
|
| 20 |
+
"epoch": 0.9989189189189189,
|
| 21 |
+
"eval_loss": 0.002777691464871168,
|
| 22 |
+
"eval_runtime": 792.7494,
|
| 23 |
+
"eval_samples_per_second": 3359.543,
|
| 24 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.975327415818089,
|
| 25 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.7701693773269653,
|
| 26 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.9911172257655966,
|
| 27 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9624723810391563,
|
| 28 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.7597355842590332,
|
| 29 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9441287064717891,
|
| 30 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9527175567355498,
|
| 31 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.9724290300680068,
|
| 32 |
+
"eval_steps_per_second": 4.375,
|
| 33 |
+
"step": 693
|
| 34 |
+
},
|
| 35 |
+
{
|
| 36 |
+
"epoch": 1.4425225225225224,
|
| 37 |
+
"grad_norm": 0.11966603249311447,
|
| 38 |
+
"learning_rate": 2.1327185244587008e-05,
|
| 39 |
+
"loss": 0.0128,
|
| 40 |
+
"step": 1000
|
| 41 |
+
},
|
| 42 |
+
{
|
| 43 |
+
"epoch": 1.998918918918919,
|
| 44 |
+
"eval_loss": 0.002118302509188652,
|
| 45 |
+
"eval_runtime": 789.0568,
|
| 46 |
+
"eval_samples_per_second": 3375.265,
|
| 47 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.9846242227629088,
|
| 48 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.6801187992095947,
|
| 49 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.9955904030209028,
|
| 50 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9765449140552956,
|
| 51 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.6780189275741577,
|
| 52 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9651303277408154,
|
| 53 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9721848413657824,
|
| 54 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.9809442711989229,
|
| 55 |
+
"eval_steps_per_second": 4.395,
|
| 56 |
+
"step": 1386
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 2.1643243243243244,
|
| 60 |
+
"grad_norm": 0.04637068510055542,
|
| 61 |
+
"learning_rate": 1.5312750601443466e-05,
|
| 62 |
+
"loss": 0.0084,
|
| 63 |
+
"step": 1500
|
| 64 |
+
},
|
| 65 |
+
{
|
| 66 |
+
"epoch": 2.885045045045045,
|
| 67 |
+
"grad_norm": 0.07652924209833145,
|
| 68 |
+
"learning_rate": 9.29831595829992e-06,
|
| 69 |
+
"loss": 0.0065,
|
| 70 |
+
"step": 2000
|
| 71 |
+
},
|
| 72 |
+
{
|
| 73 |
+
"epoch": 2.998918918918919,
|
| 74 |
+
"eval_loss": 0.0014588695485144854,
|
| 75 |
+
"eval_runtime": 782.4422,
|
| 76 |
+
"eval_samples_per_second": 3403.799,
|
| 77 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.9879171547865789,
|
| 78 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.7181636691093445,
|
| 79 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.996840725826042,
|
| 80 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9815604299892273,
|
| 81 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.7181636691093445,
|
| 82 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9725931427811844,
|
| 83 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9775832353646149,
|
| 84 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.98557011840788,
|
| 85 |
+
"eval_steps_per_second": 4.432,
|
| 86 |
+
"step": 2079
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 3.606846846846847,
|
| 90 |
+
"grad_norm": 0.04797244444489479,
|
| 91 |
+
"learning_rate": 3.2838813151563755e-06,
|
| 92 |
+
"loss": 0.0056,
|
| 93 |
+
"step": 2500
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 3.998918918918919,
|
| 97 |
+
"eval_loss": 0.001379696768708527,
|
| 98 |
+
"eval_runtime": 780.7313,
|
| 99 |
+
"eval_samples_per_second": 3411.258,
|
| 100 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy": 0.9885216725241056,
|
| 101 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_accuracy_threshold": 0.7183246612548828,
|
| 102 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_ap": 0.9971022799526896,
|
| 103 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1": 0.9824706124974221,
|
| 104 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_f1_threshold": 0.7085607051849365,
|
| 105 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_mcc": 0.9739458779668466,
|
| 106 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_precision": 0.9782229269572558,
|
| 107 |
+
"eval_sentence-transformers-paraphrase-multilingual-MiniLM-L12-v2_cosine_recall": 0.9867553479166427,
|
| 108 |
+
"eval_steps_per_second": 4.442,
|
| 109 |
+
"step": 2772
|
| 110 |
+
}
|
| 111 |
+
],
|
| 112 |
+
"logging_steps": 500,
|
| 113 |
+
"max_steps": 2772,
|
| 114 |
+
"num_input_tokens_seen": 0,
|
| 115 |
+
"num_train_epochs": 4,
|
| 116 |
+
"save_steps": 100,
|
| 117 |
+
"stateful_callbacks": {
|
| 118 |
+
"EarlyStoppingCallback": {
|
| 119 |
+
"args": {
|
| 120 |
+
"early_stopping_patience": 1,
|
| 121 |
+
"early_stopping_threshold": 0.0
|
| 122 |
+
},
|
| 123 |
+
"attributes": {
|
| 124 |
+
"early_stopping_patience_counter": 0
|
| 125 |
+
}
|
| 126 |
+
},
|
| 127 |
+
"TrainerControl": {
|
| 128 |
+
"args": {
|
| 129 |
+
"should_epoch_stop": false,
|
| 130 |
+
"should_evaluate": false,
|
| 131 |
+
"should_log": false,
|
| 132 |
+
"should_save": true,
|
| 133 |
+
"should_training_stop": true
|
| 134 |
+
},
|
| 135 |
+
"attributes": {}
|
| 136 |
+
}
|
| 137 |
+
},
|
| 138 |
+
"total_flos": 0.0,
|
| 139 |
+
"train_batch_size": 768,
|
| 140 |
+
"trial_name": null,
|
| 141 |
+
"trial_params": null
|
| 142 |
+
}
|