all-MiniLM-L6-v61-pair_score

This is a sentence-transformers model finetuned from KhaledReda/all-MiniLM-L6-v60-pair_score on the pairs_with_scores_v49 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: KhaledReda/all-MiniLM-L6-v60-pair_score
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • pairs_with_scores_v49
• Language: en
• License: apache-2.0

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, 'architecture': '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("sentence_transformers_model_id")
# Run inference
sentences = [
    'fogg master spray',
    'toledo diet tuna one piece healthy tuna toledo diet tuna one piece offers a guilt-free option for tuna lovers. packed with omega-3 fatty acids and lean protein it s a convenient option for busy days when you need a quick and healthy meal on the go.',
    'oral-b allrounder black family tooth brush 4 pieces allrounder tooth brush',
]
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, -0.1296, -0.0304],
#         [-0.1296,  1.0000, -0.0104],
#         [-0.0304, -0.0104,  1.0000]])

Training Details

Training Dataset

pairs_with_scores_v49

• Dataset: pairs_with_scores_v49 at 1cf26c2
• Size: 43,464,314 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 4 tokensmean: 7.49 tokensmax: 22 tokens	min: 3 tokensmean: 60.23 tokensmax: 242 tokens	min: 0.0mean: 0.25max: 1.0

• Samples:

  sentence1	sentence2	score
  bawys toffee v coffee	microfibre pool towel coral size l 80 x 130 cm coral towel coral pool towel lightweight pool towel ultracompact pool towel absorbent pool towel our design teams developed this towel for swimmers and all other athletes who want to dry themselves effectively. limited space in your bag take this ultra-compact and lightweight microfibre towel with you anywhere. this pool towel is super absorbent and it dries very quickly so it can be used again in no time.	0.0
  paradise - supreme mistika pistachio ice cream -	balance board the balance board is currently our only balancing that can be enjoyed by both older children and adults alike. starting 6 years old can use the board to increase their core strength and focus. yet it is advised to supervise children between 6 and 9 years old and help them find their balance in the beginning. you can do your workout indoors when the weather is simply too hot or take your board and roll to the beach. the balance board comes in different color finishes. we use wood stain to color the board to let the natural look of the wood shine. finally the board is finished with a varnish to keep the board looking nice when you choose to take it outdoors. the balance board comes	0.0
  soccer bibs	el kaff cufflinks sterling silver cufflink brass cufflink blue sapphire cufflink gold cufflink silver cufflink men cufflink 925 sterling silver and brass. plated in gold set with a blue sapphire.	0.0

• Loss: CoSENTLoss with these parameters:

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

Evaluation Dataset

pairs_with_scores_v49

• Dataset: pairs_with_scores_v49 at 1cf26c2
• Size: 218,414 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 3 tokensmean: 7.53 tokensmax: 25 tokens	min: 3 tokensmean: 59.16 tokensmax: 213 tokens	min: 0.0mean: 0.27max: 1.0

• Samples:

  sentence1	sentence2	score
  marevan warfarin sodium	retro print overshirt in blue loose fitting overshirt collared overshirt drop shoulders overshirt flap chest pockets overshirt long sleeves overshirt buttons overshirt jacket overshirt women overshirt retro overshirt poster overshirt retro poster overshirt in loose fitting collared printed jacket featuring drop shoulders flap chest pockets long sleeves and buttons down the front.	0.0
  water resistant bracelet	purple honeycomb ball add a pop of color to your next party with these purple honeycomb balls they are perfect on their own but can also be fabulous as a table centerpiece. available in different colors and sizes. includes one honeycomb ball. arrives flat. size 10 inch - 30 cm. simply assembly required. includes string for hanging. reusable.	0.0
  sunshine mackerel	cuttle fish seacube 500g seacube cuttle fish	1.0

• Loss: CoSENTLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• learning_rate: 2e-05
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: True

All Hyperparameters

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• 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: 2e-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: 1
• 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: 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: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• 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
• hub_revision: None
• 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
• liger_kernel_config: None
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Framework Versions

• Python: 3.12.3
• Sentence Transformers: 5.1.0
• Transformers: 4.55.4
• PyTorch: 2.6.0+cu124
• Accelerate: 1.10.1
• Datasets: 4.0.0
• Tokenizers: 0.21.4

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",
}

CoSENTLoss

@online{kexuefm-8847,
    title={CoSENT: A more efficient sentence vector scheme than Sentence-BERT},
    author={Su Jianlin},
    year={2022},
    month={Jan},
    url={https://kexue.fm/archives/8847},
}