PawK/GIS4MODEL

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2 on the json dataset. 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.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: sentence-transformers/all-MiniLM-L6-v2
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • json

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': 256, 'do_lower_case': False}) with Transformer model: BertModel 
  (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})
  (2): Normalize()
)

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("PawK/GIS4MODEL")
# Run inference
sentences = [
    'Threat 4 from Ball Aerospace James Webb Telescope security vulnerability Threat involving Threat 4 from Ball Aerospace James Webb Telescope security vulnerability',
    'CVE-2023-29360: SIMULATED-1 This is a simulated vulnerability for testing with keyword: Ball Aerospace James Webb Telescope security vulnerability security Yes 2025-07-01 Ball Aerospace James Webb Telescope security vulnerability Risk Score: 6.194262717622377 Severity: Critical Exploitability: Moderate Impact: Significant Affected: Tech Industries Components Manufacturers: Tech Industries',
    'CVE-2023-32315: SIMULATED-2 This is a simulated vulnerability for testing with keyword: James Webb Space Telescope compromise security No 2025-08-02 James Webb Space Telescope compromise Risk Score: 5.359379620388373 Severity: Critical Exploitability: Easy Impact: Minimal Affected: Cybersecurity Inc Components Manufacturers: Cybersecurity Inc',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

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

Training Details

Training Dataset

json

• Dataset: json
• Size: 769 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 769 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 11 tokens mean: 15.14 tokens max: 30 tokens	min: 70 tokens mean: 78.26 tokens max: 89 tokens	min: 70 tokens mean: 78.31 tokens max: 89 tokens

• Samples:

  anchor	positive	negative
  Asset related to Asset 2 from Ball Aerospace James Webb Telescope security vulnerability	CVE-2023-36036: SIMULATED-3 This is a simulated vulnerability for testing with keyword: Ball Aerospace James Webb Telescope security vulnerability security No 2025-09-05 Ball Aerospace James Webb Telescope security vulnerability Risk Score: 2.1667727327059936 Severity: High Exploitability: Moderate Impact: Minimal Affected: Generic Corp Components Manufacturers: Generic Corp	CVE-2023-29360: SIMULATED-3 This is a simulated vulnerability for testing with keyword: Webb Telescope attack security Planned 2025-07-15 Webb Telescope attack Risk Score: 7.06990194191305 Severity: Medium Exploitability: Easy Impact: Minimal Affected: Cybersecurity Inc Components Manufacturers: Cybersecurity Inc
  Asset related to Asset 3 from James Webb malware	CVE-2022-0847: SIMULATED-5 This is a simulated vulnerability for testing with keyword: James Webb malware malware Planned 2025-06-18 James Webb malware Risk Score: 3.346935150769055 Severity: Medium Exploitability: Easy Impact: Minimal Affected: Cybersecurity Inc Components Manufacturers: Cybersecurity Inc	CVE-2023-32315: SIMULATED-3 This is a simulated vulnerability for testing with keyword: James Webb CVE security No 2025-07-09 James Webb CVE Risk Score: 5.640697982524985 Severity: Low Exploitability: Moderate Impact: Minimal Affected: Tech Industries Components Manufacturers: Tech Industries
  Asset related to Asset 3 from Webb Telescope backdoor	CVE-2023-32315: SIMULATED-4 This is a simulated vulnerability for testing with keyword: Webb Telescope backdoor security Yes 2025-09-13 Webb Telescope backdoor Risk Score: 4.6647547554970785 Severity: Medium Exploitability: Difficult Impact: Significant Affected: Tech Industries Components Manufacturers: Tech Industries	CVE-2023-29360: SIMULATED-4 This is a simulated vulnerability for testing with keyword: Ball Aerospace James Webb Telescope security vulnerability security Yes 2025-09-13 Ball Aerospace James Webb Telescope security vulnerability Risk Score: 0.5088432907602923 Severity: Medium Exploitability: Difficult Impact: Critical Affected: Generic Corp Components Manufacturers: Generic Corp

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 5
  }
  

Evaluation Dataset

json

• Dataset: json
• Size: 193 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 193 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 11 tokens mean: 15.33 tokens max: 30 tokens	min: 70 tokens mean: 78.34 tokens max: 89 tokens	min: 70 tokens mean: 77.8 tokens max: 88 tokens

• Samples:

  anchor	positive	negative
  Asset related to Asset 1 from James Webb backdoor	CVE-2023-3456: SIMULATED-3 This is a simulated vulnerability for testing with keyword: James Webb backdoor security Yes 2025-07-13 James Webb backdoor Risk Score: 0.8440758738483046 Severity: Medium Exploitability: Difficult Impact: Critical Affected: Cybersecurity Inc Components Manufacturers: Cybersecurity Inc	CVE-2023-28771: SIMULATED-2 This is a simulated vulnerability for testing with keyword: Webb Telescope compromise security No 2025-06-15 Webb Telescope compromise Risk Score: 1.310545683795466 Severity: Medium Exploitability: Moderate Impact: Significant Affected: Generic Corp Components Manufacturers: Generic Corp
  Asset related to Asset 4 from James Webb breach	CVE-2023-3456: SIMULATED-3 This is a simulated vulnerability for testing with keyword: James Webb breach security No 2025-08-12 James Webb breach Risk Score: 3.4028923165168137 Severity: Medium Exploitability: Easy Impact: Critical Affected: Generic Corp Components Manufacturers: Generic Corp	CVE-2023-21036: SIMULATED-2 This is a simulated vulnerability for testing with keyword: James Webb Space Telescope security security Yes 2025-10-06 James Webb Space Telescope security Risk Score: 3.0882717202625423 Severity: Critical Exploitability: Easy Impact: Moderate Affected: Tech Industries Components Manufacturers: Tech Industries
  Asset related to Asset 3 from James Webb Space Telescope security	CVE-2023-32315: SIMULATED-1 This is a simulated vulnerability for testing with keyword: James Webb Space Telescope security security Yes 2025-09-03 James Webb Space Telescope security Risk Score: 6.962623430566551 Severity: High Exploitability: Moderate Impact: Significant Affected: Generic Corp Components Manufacturers: Generic Corp	CVE-2023-28771: SIMULATED-1 This is a simulated vulnerability for testing with keyword: James Webb Space Telescope threat security No 2025-09-23 James Webb Space Telescope threat Risk Score: 7.572013580016532 Severity: Low Exploitability: Moderate Impact: Moderate Affected: Tech Industries Components Manufacturers: Tech Industries

• Loss: TripletLoss with these parameters:

  {
      "distance_metric": "TripletDistanceMetric.COSINE",
      "triplet_margin": 5
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• num_train_epochs: 4
• warmup_ratio: 0.1
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 4
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.1
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• save_safetensors: True
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• no_cuda: False
• use_cpu: False
• use_mps_device: False
• seed: 42
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: False
• fp16: False
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: 0
• ddp_backend: None
• tpu_num_cores: None
• tpu_metrics_debug: False
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• past_index: -1
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_min_num_params: 0
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• tp_size: 0
• fsdp_transformer_layer_cls_to_wrap: None
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• skip_memory_metrics: True
• use_legacy_prediction_loop: False
• push_to_hub: False
• resume_from_checkpoint: None
• hub_model_id: None
• hub_strategy: every_save
• hub_private_repo: None
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• include_for_metrics: []
• eval_do_concat_batches: True
• fp16_backend: auto
• push_to_hub_model_id: None
• push_to_hub_organization: None
• mp_parameters:
• auto_find_batch_size: False
• full_determinism: False
• torchdynamo: None
• ray_scope: last
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• include_tokens_per_second: False
• include_num_input_tokens_seen: False
• neftune_noise_alpha: None
• optim_target_modules: None
• batch_eval_metrics: False
• eval_on_start: False
• use_liger_kernel: False
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss
2.0	50	4.0999	-
2.32	60	4.0104	-
2.7200	70	4.0787	-
3.12	80	3.9972	-
3.52	90	3.9994	-
3.92	100	4.0216	3.8883

Framework Versions

• Python: 3.13.3
• Sentence Transformers: 4.1.0
• Transformers: 4.51.3
• PyTorch: 2.6.0+cpu
• Accelerate: 1.6.0
• 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",
}

TripletLoss

@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification},
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}