SentenceTransformer based on intfloat/multilingual-e5-small

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-small. It maps sentences & paragraphs to a 384-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, classification, clustering, and more.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: intfloat/multilingual-e5-small
Maximum Sequence Length: 512 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 = [
    'query: psychology of desire and anger',
    'passage: ध्यायतो विषयान्पुंसः सङ्गस्तेषूपजायते। सङ्गात्सञ्जायते कामः कामात्क्रोधोऽभिजायते॥ (dhyayato visayan pumsah sangas tesupajayate) — Dwelling on sense objects creates attachment; attachment breeds desire; and desire, when frustrated, gives rise to anger.',
    'passage: अनन्याश्चिन्तयन्तो मां ये जनाः पर्युपासते। तेषां नित्याभियुक्तानां योगक्षेमं वहाम्यहम्॥ (ananyas cintayanto mam) — For those who worship with single-minded devotion, thinking of nothing else, I personally provide for what they lack and preserve what they have.',
]
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, 0.8061, 0.4757],
#         [0.8061, 1.0000, 0.4585],
#         [0.4757, 0.4585, 1.0000]])

Training Details

Training Dataset

Unnamed Dataset

Size: 52 training samples
Columns: sentence_0, sentence_1, and sentence_2

Approximate statistics based on the first 52 samples:

	sentence_0	sentence_1	sentence_2
type	string	string	string
modality	text	text	text
details	min: 9 tokens mean: 20.21 tokens max: 50 tokens	min: 75 tokens mean: 87.38 tokens max: 100 tokens	min: 75 tokens mean: 88.17 tokens max: 100 tokens

Samples:

sentence_0	sentence_1	sentence_2
`query: psychology of desire and anger`	`passage: ध्यायतो विषयान्पुंसः सङ्गस्तेषूपजायते। सङ्गात्सञ्जायते कामः कामात्क्रोधोऽभिजायते॥ (dhyayato visayan pumsah sangas tesupajayate) — Dwelling on sense objects creates attachment; attachment breeds desire; and desire, when frustrated, gives rise to anger.`	`passage: अनन्याश्चिन्तयन्तो मां ये जनाः पर्युपासते। तेषां नित्याभियुक्तानां योगक्षेमं वहाम्यहम्॥ (ananyas cintayanto mam) — For those who worship with single-minded devotion, thinking of nothing else, I personally provide for what they lack and preserve what they have.`
`query: The self is never born and never dies; it is unborn, eternal, permanent and ancient, and is not slain when the body is slain.`	`passage: न जायते म्रियते वा कदाचिन्नायं भूत्वा भविता वा न भूयः। अजो नित्यः शाश्वतोऽयं पुराणो न हन्यते हन्यमाने शरीरे॥ (na jayate mriyate va kadacin) — The self is never born and never dies; it is unborn, eternal, permanent and ancient, and is not slain when the body is slain.`	`passage: मात्रास्पर्शास्तु कौन्तेय शीतोष्णसुखदुःखदाः। आगमापायिनोऽनित्यास्तांस्तितिक्षस्व भारत॥ (matra-sparsas tu kaunteya) — Contact between the senses and their objects gives rise to heat and cold, pleasure and pain; these are fleeting and impermanent, so one should learn to tolerate them.`
`query: why follow your own path in life`	`passage: श्रेयान्स्वधर्मो विगुणः परधर्मात्स्वनुष्ठितात्। स्वधर्मे निधनं श्रेयः परधर्मो भयावहः॥ (sreyan sva-dharmo vigunah) — It is better to imperfectly follow one's own path than to perfectly follow another's; better to die pursuing one's own duty, for another's path is fraught with danger.`	`passage: कर्मण्यकर्म यः पश्येदकर्मणि च कर्म यः। स बुद्धिमान्मनुष्येषु स युक्तः कृत्स्नकर्मकृत्॥ (karmany akarma yah pasyed) — One who sees inaction in action, and action in inaction, is wise among people; such a person, though engaged in all activities, remains a true practitioner of yoga.`

Loss: MultipleNegativesRankingLoss with these parameters:

{
    "scale": 20.0,
    "similarity_fct": "cos_sim",
    "gather_across_devices": false,
    "directions": [
        "query_to_doc"
    ],
    "partition_mode": "joint",
    "hardness_mode": null,
    "hardness_strength": 0.0
}

Training Hyperparameters

Non-Default Hyperparameters

num_train_epochs: 6
fp16: True
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

per_device_train_batch_size: 8
num_train_epochs: 6
max_steps: -1
learning_rate: 5e-05
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_steps: 0
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
label_smoothing_factor: 0.0
bf16: False
fp16: True
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
dataloader_multiprocessing_context: None
dataloader_in_order: True
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: None
fsdp_config: None
deepspeed: None
debug: []
skip_memory_metrics: True
do_predict: False
resume_from_checkpoint: None
local_rank: -1
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin
router_mapping: {}
learning_rate_mapping: {}
warmup_ratio: None

Training Time

Training: 4.0 minutes

Framework Versions

Python: 3.13.15
Sentence Transformers: 5.7.0
Transformers: 5.15.0
PyTorch: 2.11.0+cu128
Accelerate: 1.14.0
Datasets: 4.0.0
Tokenizers: 0.22.2

Additional Resources

Training and Finetuning Embedding Models with Sentence Transformers: the end-to-end guide for training or finetuning Sentence Transformer models.
Introduction to Matryoshka Embedding Models: variable-size embeddings that can be truncated with minimal quality loss.
Binary and Scalar Embedding Quantization for Significantly Faster & Cheaper Retrieval: post-training compression of embedding vectors.
Multimodal Embedding & Reranker Models with Sentence Transformers: use text, image, audio, and video models through the same API.
Training and Finetuning Multimodal Embedding & Reranker Models with Sentence Transformers: train multimodal embedding models, with a Visual Document Retrieval walkthrough.

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{oord2019representationlearningcontrastivepredictive,
      title={Representation Learning with Contrastive Predictive Coding},
      author={Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      year={2019},
      eprint={1807.03748},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/1807.03748},
}