Shuu12121 / CodeSearch-ModernBERT-Snake-Plus

SentenceTransformer based on Shuu12121/CodeModernBERT-Snake

This is a sentence-transformers model finetuned from Shuu12121/CodeModernBERT-Snake . It maps sentences & paragraphs to a 512-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: Shuu12121/CodeModernBERT-Snake
• Maximum Sequence Length: 1024 tokens
• Output Dimensionality: 512 dimensions
• Similarity Function: Cosine Similarity

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 1024, 'do_lower_case': False}) with Transformer model: ModernBertModel 
  (1): Pooling({'word_embedding_dimension': 512, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("sentence_transformers_model_id")
# Run inference
sentences = [
    'This method calculates the turn weight separately.',
    'public double calcTurnWeight(int edgeFrom, int nodeVia, int edgeTo) {\n        long turnFlags = turnCostExt.getTurnCostFlags(edgeFrom, nodeVia, edgeTo);\n        if (turnCostEncoder.isTurnRestricted(turnFlags))\n            return Double.POSITIVE_INFINITY;\n\n        return turnCostEncoder.getTurnCost(turnFlags);\n    }',
    'public void putAllWriteable(BeanMap<T> map) {\n        map.types.keySet().stream().filter(key -> getWriteInvoker(key) != null).forEach(key -> this.put(key, map.get(key)));\n    }',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 512]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Training Details

Training Dataset

Unnamed Dataset

• Size: 1,761,750 training samples
• Columns: sentence_0 , sentence_1 , and label
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	label
  type	string	string	float
  details	min: 3 tokens mean: 47.87 tokens max: 633 tokens	min: 28 tokens mean: 164.44 tokens max: 1024 tokens	min: 1.0 mean: 1.0 max: 1.0

• Samples:

  sentence_0	sentence_1	label
  // Read reads from serial port. // It is blocked until data received or timeout after p.timeout.	func (p *port) Read(b []byte) (n int, err error) { var done uint32 if err = syscall.ReadFile(p.handle, b, &done, nil); err != nil { return } if done == 0 { err = ErrTimeout return } n = int(done) return }	1.0
  // _NET_WM_STRUT_PARTIAL set	func WmStrutPartialSet(xu *xgbutil.XUtil, win xproto.Window, struts *WmStrutPartial) error { rawStruts := make([]uint, 12) rawStruts[0] = struts.Left rawStruts[1] = struts.Right rawStruts[2] = struts.Top rawStruts[3] = struts.Bottom rawStruts[4] = struts.LeftStartY rawStruts[5] = struts.LeftEndY rawStruts[6] = struts.RightStartY rawStruts[7] = struts.RightEndY rawStruts[8] = struts.TopStartX rawStruts[9] = struts.TopEndX rawStruts[10] = struts.BottomStartX rawStruts[11] = struts.BottomEndX return xprop.ChangeProp32(xu, win, "_NET_WM_STRUT_PARTIAL", "CARDINAL", rawStruts...) }	1.0
  // Union returns a new geometry representing all points in this geometry and the // other.	func (g *Geometry) Union(other *Geometry) (*Geometry, error) { return g.binaryTopo("Union", cGEOSUnion, other) }	1.0

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size : 400
• per_device_eval_batch_size : 400
• num_train_epochs : 5
• fp16 : True
• multi_dataset_batch_sampler : round_robin

All Hyperparameters

Training Logs

Framework Versions

• Python: 3.11.12
• Sentence Transformers: 3.4.1
• Transformers: 4.51.3
• PyTorch: 2.6.0+cu124
• Accelerate: 1.5.2
• Datasets: 3.5.0
• Tokenizers: 0.21.1

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}