SentenceTransformer based on jhu-clsp/mmBERT-base

This is a sentence-transformers model finetuned from jhu-clsp/mmBERT-base. It maps sentences & paragraphs to a 768-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: jhu-clsp/mmBERT-base
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity

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': 'ModernBertModel'})
  (1): Pooling({'word_embedding_dimension': 768, '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})
)

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("yjoonjang/MIMO-mmBERT-base")
# Run inference
sentences = [
    'Third Ave.',
    'Third Ave.',
    'it',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 1.0000, 0.4479],
#         [1.0000, 1.0000, 0.4479],
#         [0.4479, 0.4479, 1.0000]])

Evaluation

Metrics

Information Retrieval

• Datasets: NanoMIRACL-ar, NanoMIRACL-de, NanoMIRACL-en, NanoMIRACL-es, NanoMIRACL-fr, NanoMIRACL-hi, NanoMIRACL-id, NanoMIRACL-ja, NanoMIRACL-ru and NanoMIRACL-zh
• Evaluated with InformationRetrievalEvaluator

Nano MIRACL

• Dataset: NanoMIRACL_mean
• Evaluated with evaluation.nano_miracl_evaluator.NanoMIRACLEvaluator

Training Details

Training Dataset

Unnamed Dataset

• Size: 5,647,936 training samples
• Columns: anchor, positive, and lang
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	lang
  type	string	string	string
  details	min: 4 tokensmean: 26.92 tokensmax: 129 tokens	min: 4 tokensmean: 22.73 tokensmax: 118 tokens	min: 3 tokensmean: 3.0 tokensmax: 3 tokens

• Samples:

  anchor	positive	lang
  Desidero soltanto far presente che le due proposte presentate dal nostro gruppo sulla questione dei trasferimenti di crediti non sono risultate del tutto compatibili e pertanto sono state respinte.	I only wish to draw attention to the two amendments tabled by our group on the question of carry-overs, where the texts are incompatible in their present form. They were rejected.	it
  Dieser ist - so schwierig es ist, die Klimaveränderung mit Rechenexempeln zu demonstrieren - zum Teil von Menschen gemacht.	This is partly the work of human hand, as difficult as it is to demonstrate climate change using calculations.	de
  Nel corso della procedura a) il Parlamento sarà consultato in merito al regime linguistico (la decisione in questo caso deve essere assunta dal Consiglio), b) il Parlamento prenderà parte al processo decisionale sul contenuto del regolamento sul brevetto nell'ambito della procedura legislativa ordinaria, c) è previsto il consenso dell'Assemblea sulla giurisdizione del brevetto.	The remainder of the procedure will involve a) Parliament being consulted about the language arrangements (the decision in this case must be made by the Council), b) Parliament taking part in deciding on the content of the patent regulation as part of the ordinary legislative procedure, c) Parliament's consent being obtained for the patent jurisdiction. There will be no change in any of these opportunities for Parliament to participate in the process.	it

• Loss: custom_losses.embed_distill_loss.EmbedDistillLoss with these parameters:

  {
      "distance_metric": "cosine",
      "projection_in": 768,
      "projection_out": 4096
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 128
• num_train_epochs: 1.0
• learning_rate: 0.0001
• warmup_steps: 0.1
• bf16: True
• eval_strategy: steps
• eval_on_start: True
• dataloader_num_workers: 4
• warmup_ratio: 0.1

All Hyperparameters

• per_device_train_batch_size: 128
• num_train_epochs: 1.0
• max_steps: -1
• learning_rate: 0.0001
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• warmup_steps: 0.1
• optim: adamw_torch_fused
• optim_args: None
• weight_decay: 0.0
• 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: True
• 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: trackio
• eval_strategy: steps
• per_device_eval_batch_size: 8
• prediction_loss_only: True
• eval_on_start: True
• 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: True
• dataloader_num_workers: 4
• 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_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: 0.1
• local_rank: -1
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}
• max_seq_length: 256
• stage: 1
• distance_metric: cosine
• infonce_weight: 0.5
• distill_weight: 0.5
• infonce_scale: 20.0
• mini_batch_size: 32
• teacher_query_prompt: None
• teacher_max_seq_length: 256
• projection_path: None
• group_by_language: False
• pooling_mode: mean
• student_query_prompt: None
• student_doc_prompt: None

Training Logs

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 5.2.3
• Transformers: 5.2.0
• PyTorch: 2.8.0+cu128
• Accelerate: 1.12.0
• Datasets: 4.6.1
• 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",
}