SentenceTransformer based on intfloat/multilingual-e5-large

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-large on an augmented version of stsb_multi_es dataset. It maps sentences & paragraphs to a 1024-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: intfloat/multilingual-e5-large
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 1024 tokens
• Similarity Function: Cosine Similarity
• Training Dataset:
  • stsb_multi_es_aug

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': 512, 'do_lower_case': False}) with Transformer model: XLMRobertaModel 
  (1): Pooling({'word_embedding_dimension': 1024, '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("mrm8488/multilingual-e5-large-ft-sts-spanish-matryoshka-768-16-5e")
# Run inference
sentences = [
    'El avión está tocando tierra.',
    'El avión animado se encuentra en proceso de aterrizaje.',
    'Un pequeño niño montado en un columpio en el parque.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

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

Evaluation

Metrics

Semantic Similarity

• Dataset: sts-dev-768
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-dev-512
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-dev-256
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-dev-128
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-dev-64
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-dev-32
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-dev-16
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-test-768
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-test-512
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-test-256
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-test-128
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-test-64
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-test-32
• Evaluated with EmbeddingSimilarityEvaluator

Semantic Similarity

• Dataset: sts-test-16
• Evaluated with EmbeddingSimilarityEvaluator

Training Details

Training Dataset

stsb_multi_es_aug

• Dataset: stsb_multi_es_aug
• Size: 2,697 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 8 tokensmean: 22.25 tokensmax: 68 tokens	min: 8 tokensmean: 22.01 tokensmax: 79 tokens	min: 0.0mean: 2.67max: 5.0

• Samples:

  sentence1	sentence2	score
  El pájaro de tamaño reducido se posó con delicadeza en una rama cubierta de escarcha.	Un ave de color amarillo descansaba tranquilamente en una rama.	3.200000047683716
  Una chica está tocando la flauta en un parque.	Un grupo de músicos está tocando en un escenario al aire libre.	1.286
  La aclamada escritora británica, Doris Lessing, galardonada con el premio Nobel, fallece	La destacada autora británica, Doris Lessing, reconocida con el prestigioso Premio Nobel, muere	4.199999809265137

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "CoSENTLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64,
          32,
          16
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Evaluation Dataset

stsb_multi_es_aug

• Dataset: stsb_multi_es_aug
• Size: 697 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 8 tokensmean: 22.76 tokensmax: 67 tokens	min: 7 tokensmean: 22.26 tokensmax: 63 tokens	min: 0.0mean: 2.3max: 5.0

• Samples:

  sentence1	sentence2	score
  Un incendio ocurrido en un hospital psiquiátrico ruso resultó en la trágica muerte de 38 personas.	Se teme que el incendio en un hospital psiquiátrico ruso cause la pérdida de la vida de 38 individuos.	4.199999809265137
  "Street dijo que el otro individuo a veces se siente avergonzado de su fiesta, lo cual provoca risas en la multitud"	"A veces, el otro tipo se encuentra avergonzado de su fiesta y no se le puede culpar."	3.5
  El veterano diplomático de Malasia tuvo un encuentro con Suu Kyi el miércoles en la casa del lago en Yangon donde permanece bajo arresto domiciliario.	Razali Ismail tuvo una reunión de 90 minutos con Suu Kyi, quien ganó el Premio Nobel de la Paz en 1991, en su casa del lago donde está recluida.	3.691999912261963

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "CoSENTLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          128,
          64,
          32,
          16
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• num_train_epochs: 5
• 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: 16
• per_device_eval_batch_size: 16
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_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: 5
• 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: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• 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
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.0.0
• Transformers: 4.41.1
• PyTorch: 2.3.0+cu121
• Accelerate: 0.30.1
• Datasets: 2.19.1
• Tokenizers: 0.19.1

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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning}, 
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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},
}