tjohn327/scion-all-MiniLM-L6-v2

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

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-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.

Model Details

Model Description

• —Model Type: Sentence Transformer
• —Base model: sentence-transformers/all-MiniLM-L6-v2 <!-- at revision fa97f6e7cb1a59073dff9e6b13e2715cf7475ac9 -->
• —Maximum Sequence Length: 256 tokens
• —Output Dimensionality: 384 dimensions
• —Similarity Function: Cosine Similarity <!-- - Training Dataset: Unknown --> <!-- - Language: Unknown --> <!-- - License: Unknown -->

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("tjohn327/scion-all-MiniLM-L6-v2")
# Run inference
sentences = [
    "How does SCION's time synchronization support duplicate detection and traffic monitoring?",
    'The document chunk details the COLIBRI system for bandwidth reservations within the SCION architecture, emphasizing its efficiency in managing end-to-end reservations (EERs) across on-path Autonomous Systems (ASes). COLIBRI employs symmetric cryptography for stateless verification of reservations, mitigating the need for per-source state. To counter replay attacks, a duplicate suppression mechanism is essential, ensuring that authenticated packets cannot be maliciously reused to exceed allocated bandwidth. Monitoring and policing systems are necessary to enforce bandwidth limits, allowing ASes to identify and manage misbehaving flows. While stateful flow monitoring, such as the token-bucket algorithm, provides precise measurements, it is limited to edge deployments due to state requirements. In contrast, probabilistic monitoring techniques like LOFT can operate effectively within the Internet core, accommodating a vast number of flows. Time synchronization is critical for coordinating bandwidth reservations across AS boundaries, facilitating accurate duplicate detection and traffic monitoring. SCION’s time synchronization mechanism ensures adequate synchronization levels for these operations. Overall, COLIBRI enables efficient short-term bandwidth reservations, crucial for maintaining the integrity of end-to-end communications in a highly dynamic Internet environment.',
    "SCION is a next-generation Internet architecture designed to enhance security, availability, isolation, and scalability. It enables end hosts to utilize multiple authenticated inter-domain paths to any destination, with each packet carrying its own specified forwarding path determined by the sender. SCION architecture incorporates isolation domains (ISDs), which consist of multiple Autonomous Systems (ASes) that agree on a trust root configuration (TRC) defining the roots of trust for validating bindings between names and public keys or addresses. Core ASes govern each ISD, providing inter-ISD connectivity and managing trust roots. The architecture delineates three types of AS relationships: core, provider-customer, and peer-peer, with core relations existing solely among core ASes. SCION addressing is structured as a tuple ⟨ISD number, AS number, local address⟩, where the ISD number identifies the ISD of the end host, the AS number identifies the host's AS, and the local address serves as the host's identifier within its AS, not utilized for inter-domain routing or forwarding. Path-construction, path-registration, and path-lookup procedures are integral to SCION's operational framework, facilitating efficient packet forwarding based on the specified paths.",
]
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]

<!--

Direct Usage (Transformers)

<details><summary>Click to see the direct usage in Transformers</summary>

</details> -->

<!--

Downstream Usage (Sentence Transformers)

You can finetune this model on your own dataset.

<details><summary>Click to expand</summary>

</details> -->

<!--

Out-of-Scope Use

List how the model may foreseeably be misused and address what users ought not to do with the model. -->

Evaluation

Metrics

Information Retrieval

• —Dataset: dev-evaluation
• —Evaluated with <code>InformationRetrievalEvaluator</code>

<!--

Bias, Risks and Limitations

What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. -->

<!--

Recommendations

What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. -->

Training Details

Training Dataset

Unnamed Dataset

• —Size: 21,376 training samples
• —Columns: <code>sentence0</code>, <code>sentence1</code>, and <code>label</code>
• —Approximate statistics based on the first 1000 samples: | | sentence0 | sentence1 | label | |:--------|:----------------------------------------------------------------------------------|:------------------------------------------------------------------------------------|:--------------------------------------------------------------| | type | string | string | float | | details | <ul><li>min: 7 tokens</li><li>mean: 17.45 tokens</li><li>max: 30 tokens</li></ul> | <ul><li>min: 4 tokens</li><li>mean: 231.74 tokens</li><li>max: 256 tokens</li></ul> | <ul><li>min: 1.0</li><li>mean: 1.0</li><li>max: 1.0</li></ul> |
• —Samples: | sentence0 | sentence1 | label | |:-----------------------------------------------------------------------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:-----------------| | <code>What are the advantages of using flow-volume targets over cash transfers in SCION agreements?</code> | <code>The document discusses optimization strategies for interconnection agreements in SCION, focusing on flow-volume targets and cash compensation mechanisms. The optimization problem involves determining the maximum customer demand for new path segments, denoted as ∆fmax P, and adjusting flow allowances f(a) and additional traffic ∆f(a) accordingly. Flow-volume targets provide predictability and enforceable limits, enhancing the likelihood of positive utility outcomes. Conversely, cash compensation agreements, defined by a payment π between Autonomous Systems (ASes), offer flexibility and can be concluded even when flow-volume targets yield zero solutions. The Nash Bargaining Solution is employed to determine the cash transfer that balances utilities uD(a) and uE(a) of the negotiating parties. Both methods face challenges due to private information regarding costs and pricing, which can lead to inefficiencies in negotiations. The document introduces a mechanism-assisted negotiation approac...</code> | <code>1.0</code> | | <code>What is the significance of the variable 'auth⇄a' in packet forwarding?</code> | <code>The document chunk presents a formal model for secure packet forwarding protocols within the SCION architecture, detailing the dispatch and handling of packets across internal and external channels. The functions dispatch-inta, dispatch-intc, dispatch-exta, and dispatch-extc manage the addition of packets to internal and external send queues based on their authorization status and historical context. The model defines packet structures (PKTa) that encapsulate the future path (fut), past path (past), and historical path (hist), with authorization checks against the set of authorized paths (autha) and their fragment closures (auth⇄a). The environment parameters include the set of compromised nodes (Nattr), which influences the security model by introducing potential adversarial actions. The state representation employs asynchronous channels for packet transmission, with the initial state having all channels empty. The model emphasizes the importance of maintaining i...</code> | <code>1.0</code> | | <code>How can SNC and COLIBRI SegRs be combined to enhance DRKey bootstrapping security?</code> | <code>The document chunk discusses the Protected DRKey Bootstrapping mechanism within the SCION architecture, emphasizing the integration of Service Network Controllers (SNC) and COLIBRI Segments (SegRs) to enhance the security of DRKey bootstrapping. This approach mitigates the denial-of-DRKey attack surface by ensuring that pre-shared DRKeys are securely established. The adversary model considers off-path adversaries capable of modifying, dropping, or injecting packets, while maintaining that the reservation path remains uncontested by adversaries. The document highlights the necessity of path stability in the underlying network architecture for effective bandwidth-reservation protocols. Various queuing disciplines, such as distributed weighted fair queuing (DWFQ) and rate-controlled priority queuing (PQ), are discussed for traffic isolation, with specific mechanisms to prevent starvation of lower-priority traffic. Control-plane traffic is rate-limited at infrastructure services and border...</code> | <code>1.0</code> |
• —Loss: <code>MultipleNegativesRankingLoss</code> with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• —eval_strategy: steps
• —per_device_train_batch_size: 128
• —per_device_eval_batch_size: 128
• —num_train_epochs: 2
• —fp16: True
• —multi_dataset_batch_sampler: round_robin

All Hyperparameters

<details><summary>Click to expand</summary>

• —overwrite_output_dir: False
• —do_predict: False
• —eval_strategy: steps
• —prediction_loss_only: True
• —per_device_train_batch_size: 128
• —per_device_eval_batch_size: 128
• —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
• —num_train_epochs: 2
• —max_steps: -1
• —lr_scheduler_type: linear
• —lr_scheduler_kwargs: {}
• —warmup_ratio: 0.0
• —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: True
• —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: {'minnumparams': 0, 'xla': False, 'xlafsdpv2': False, 'xlafsdpgrad_ckpt': False}
• —fsdp_transformer_layer_cls_to_wrap: None
• —accelerator_config: {'splitbatches': False, 'dispatchbatches': None, 'evenbatches': True, 'useseedablesampler': True, 'nonblocking': False, 'gradientaccumulationkwargs': 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
• —dispatch_batches: None
• —split_batches: 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: batch_sampler
• —multi_dataset_batch_sampler: round_robin

</details>

Training Logs

Framework Versions

• —Python: 3.12.3
• —Sentence Transformers: 3.4.1
• —Transformers: 4.49.0
• —PyTorch: 2.6.0+cu124
• —Accelerate: 1.4.0
• —Datasets: 3.3.2
• —Tokenizers: 0.21.0

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}
}

<!--

Glossary

Clearly define terms in order to be accessible across audiences. -->

<!--

Model Card Authors

Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. -->

<!--

Model Card Contact

Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->