Update README.md
Browse files
README.md
CHANGED
|
@@ -7,7 +7,7 @@ tags:
|
|
| 7 |
|
| 8 |
---
|
| 9 |
|
| 10 |
-
# {
|
| 11 |
|
| 12 |
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
|
| 13 |
|
|
@@ -24,57 +24,11 @@ pip install -U sentence-transformers
|
|
| 24 |
Then you can use the model like this:
|
| 25 |
|
| 26 |
```python
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
model = SentenceTransformer('{MODEL_NAME}')
|
| 31 |
-
embeddings = model.encode(sentences)
|
| 32 |
-
print(embeddings)
|
| 33 |
-
```
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
## Evaluation Results
|
| 38 |
-
|
| 39 |
-
<!--- Describe how your model was evaluated -->
|
| 40 |
-
|
| 41 |
-
For an automated evaluation of this model, see the *Sentence Embeddings Benchmark*: [https://seb.sbert.net](https://seb.sbert.net?model_name={MODEL_NAME})
|
| 42 |
-
|
| 43 |
-
|
| 44 |
-
## Training
|
| 45 |
-
The model was trained with the parameters:
|
| 46 |
-
|
| 47 |
-
**DataLoader**:
|
| 48 |
-
|
| 49 |
-
`torch.utils.data.dataloader.DataLoader` of length 45861 with parameters:
|
| 50 |
-
```
|
| 51 |
-
{'batch_size': 4, 'sampler': 'torch.utils.data.sampler.RandomSampler', 'batch_sampler': 'torch.utils.data.sampler.BatchSampler'}
|
| 52 |
```
|
| 53 |
|
| 54 |
-
**Loss**:
|
| 55 |
-
|
| 56 |
-
`sentence_transformers.losses.MultipleNegativesSymmetricRankingLoss.MultipleNegativesSymmetricRankingLoss` with parameters:
|
| 57 |
-
```
|
| 58 |
-
{'scale': 20.0, 'similarity_fct': 'cos_sim'}
|
| 59 |
-
```
|
| 60 |
-
|
| 61 |
-
Parameters of the fit()-Method:
|
| 62 |
-
```
|
| 63 |
-
{
|
| 64 |
-
"epochs": 1,
|
| 65 |
-
"evaluation_steps": 4586,
|
| 66 |
-
"evaluator": "sentence_transformers.evaluation.AlignmentandUniformityEvaluator.AlignmentandUniformityEvaluator",
|
| 67 |
-
"max_grad_norm": 1,
|
| 68 |
-
"optimizer_class": "<class 'torch.optim.adamw.AdamW'>",
|
| 69 |
-
"optimizer_params": {
|
| 70 |
-
"lr": 5e-05
|
| 71 |
-
},
|
| 72 |
-
"scheduler": "WarmupLinear",
|
| 73 |
-
"steps_per_epoch": null,
|
| 74 |
-
"warmup_steps": 4587,
|
| 75 |
-
"weight_decay": 0.01
|
| 76 |
-
}
|
| 77 |
-
```
|
| 78 |
|
| 79 |
|
| 80 |
## Full Model Architecture
|
|
@@ -90,4 +44,12 @@ SentenceTransformer(
|
|
| 90 |
|
| 91 |
## Citing & Authors
|
| 92 |
|
| 93 |
-
<!--- Describe where people can find more information -->
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 7 |
|
| 8 |
---
|
| 9 |
|
| 10 |
+
# {CLFE(ConMath)}
|
| 11 |
|
| 12 |
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 768 dimensional dense vector space and can be used for tasks like clustering or semantic search.
|
| 13 |
|
|
|
|
| 24 |
Then you can use the model like this:
|
| 25 |
|
| 26 |
```python
|
| 27 |
+
embedding_latex = model.encode([{'latex': latex}])
|
| 28 |
+
embedding_pmml = model.encode([{'mathml': pmml}])
|
| 29 |
+
embedding_cmml = model.encode([{'mathml': cmml}])
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
```
|
| 31 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 32 |
|
| 33 |
|
| 34 |
## Full Model Architecture
|
|
|
|
| 44 |
|
| 45 |
## Citing & Authors
|
| 46 |
|
| 47 |
+
<!--- Describe where people can find more information -->
|
| 48 |
+
@inproceedings{wang2023math,
|
| 49 |
+
title={Math Information Retrieval with Contrastive Learning of Formula Embeddings},
|
| 50 |
+
author={Wang, Jingyi and Tian, Xuedong},
|
| 51 |
+
booktitle={International Conference on Web Information Systems Engineering},
|
| 52 |
+
pages={97--107},
|
| 53 |
+
year={2023},
|
| 54 |
+
organization={Springer}
|
| 55 |
+
}
|