cnuland/semantic-claw-router-embeddings-v1

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 retrieval.

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
• Supported Modality: Text

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'BertModel'})
  (1): Pooling({'embedding_dimension': 384, 'pooling_mode': 'mean', '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("sentence_transformers_model_id")
# Run inference
sentences = [
    'What is 100 degrees Fahrenheit converted to Celsius?',
    'Architect a globally distributed microservices deployment platform enforcing zero-trust networking, canary releases with automated rollback based on multi-region SLOs, cross-cloud secrets rotation, and infrastructure cost allocation across AWS, Azure, and GCP.',
    'What is the default network port for HTTP?',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[ 1.0000, -0.2282,  0.9992],
#         [-0.2282,  1.0000, -0.2242],
#         [ 0.9992, -0.2242,  1.0000]])

Training Details

Training Dataset

Unnamed Dataset

• Size: 871 training samples
• Columns: sentence and label
• Approximate statistics based on the first 871 samples:

  	sentence	label
  type	string	int
  details	min: 6 tokens mean: 34.77 tokens max: 179 tokens	0: ~19.86% 1: ~28.47% 2: ~27.10% 3: ~24.57%

• Samples:

  sentence	label
  Write a discharge instruction template for patients recovering from total knee replacement surgery, including medication guidelines, physical therapy milestones, and red-flag symptoms.	1
  Prove that alpha-beta pruning in game tree search returns identical minimax values as full search for deterministic two-player zero-sum games with perfect information.	3
  Explain the zone of proximal development using a specific classroom learning scenario.	1

• Loss: BatchAllTripletLoss with these parameters:

  {
      "distance_metric": "euclidean_distance",
      "margin": 5
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 32
• num_train_epochs: 20
• learning_rate: 2e-05
• warmup_steps: 0.1
• weight_decay: 0.01
• batch_sampler: group_by_label

All Hyperparameters

Click to expand

• per_device_train_batch_size: 32
• num_train_epochs: 20
• max_steps: -1
• learning_rate: 2e-05
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• warmup_steps: 0.1
• optim: adamw_torch_fused
• optim_args: None
• weight_decay: 0.01
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• optim_target_modules: None
• gradient_accumulation_steps: 1
• average_tokens_across_devices: True
• max_grad_norm: 1.0
• label_smoothing_factor: 0.0
• bf16: False
• fp16: False
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• use_liger_kernel: False
• liger_kernel_config: None
• use_cache: False
• neftune_noise_alpha: None
• torch_empty_cache_steps: None
• auto_find_batch_size: False
• log_on_each_node: True
• logging_nan_inf_filter: True
• include_num_input_tokens_seen: no
• log_level: passive
• log_level_replica: warning
• disable_tqdm: False
• project: huggingface
• trackio_space_id: None
• trackio_bucket_id: None
• trackio_static_space_id: None
• per_device_eval_batch_size: 8
• prediction_loss_only: True
• eval_on_start: False
• eval_do_concat_batches: True
• eval_use_gather_object: False
• eval_accumulation_steps: None
• include_for_metrics: []
• batch_eval_metrics: False
• save_only_model: False
• save_on_each_node: False
• enable_jit_checkpoint: False
• push_to_hub: False
• hub_private_repo: None
• hub_model_id: None
• hub_strategy: every_save
• hub_always_push: False
• hub_revision: None
• load_best_model_at_end: False
• ignore_data_skip: False
• restore_callback_states_from_checkpoint: False
• full_determinism: False
• seed: 42
• data_seed: None
• use_cpu: False
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• parallelism_config: None
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_pin_memory: True
• dataloader_persistent_workers: False
• dataloader_prefetch_factor: None
• remove_unused_columns: True
• label_names: None
• train_sampling_strategy: random
• length_column_name: length
• ddp_find_unused_parameters: None
• ddp_bucket_cap_mb: None
• ddp_broadcast_buffers: False
• ddp_static_graph: None
• ddp_backend: None
• ddp_timeout: 1800
• fsdp: []
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• deepspeed: None
• debug: []
• skip_memory_metrics: True
• do_predict: False
• resume_from_checkpoint: None
• warmup_ratio: None
• local_rank: -1
• prompts: None
• batch_sampler: group_by_label
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss
0.3704	10	4.9309
0.7407	20	4.9160
1.1111	30	4.8961
1.4815	40	4.7981
1.8519	50	4.6499
2.2222	60	4.4857
2.5926	70	4.2761
2.9630	80	4.1235
3.3333	90	4.0387
3.7037	100	3.9355
4.0741	110	3.9115
4.4444	120	3.8240
4.8148	130	3.7550
5.1852	140	3.7214
5.5556	150	3.6528
5.9259	160	3.6749
6.2963	170	3.6371
6.6667	180	3.6124
7.0370	190	3.5848
7.4074	200	3.5935
7.7778	210	3.5504
8.1481	220	3.5539
8.5185	230	3.5427
8.8889	240	3.5122
9.2593	250	3.4894
9.6296	260	3.5147
10.0	270	3.4909
10.3704	280	3.5146
10.7407	290	3.4988
11.1111	300	3.4591
11.4815	310	3.4997
11.8519	320	3.4847
12.2222	330	3.4704
12.5926	340	3.4826
12.9630	350	3.4474
13.3333	360	3.4807
13.7037	370	3.4772
14.0741	380	3.4373
14.4444	390	3.4747
14.8148	400	3.4712
15.1852	410	3.4310
15.5556	420	3.4720
15.9259	430	3.4650
16.2963	440	3.4394
16.6667	450	3.4700
17.0370	460	3.4330
17.4074	470	3.4684
17.7778	480	3.4706
18.1481	490	3.4456
18.5185	500	3.4681
18.8889	510	3.4560
19.2593	520	3.4415
19.6296	530	3.4683
20.0	540	3.4167

Training Time

• Training: 12.6 minutes

Framework Versions

• Python: 3.11.13
• Sentence Transformers: 5.4.1
• Transformers: 5.8.0
• PyTorch: 2.11.0+cpu
• Accelerate: 1.13.0
• Datasets: 4.8.5
• Tokenizers: 0.22.2

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",
}

BatchAllTripletLoss

@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}
}