Maksim-KOS/multilingual-e5-large-instruct-saturn-planet

SentenceTransformer based on intfloat/multilingual-e5-large-instruct

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-large-instruct. 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-instruct
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 1024 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': 512, 'do_lower_case': False, 'architecture': '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("Maksim-KOS/multilingual-e5-large-instruct-saturn-planet")
# Run inference
sentences = [
    'Порог дверной ламинированный 74x968 мм серый Olovi',
    'Порог дверной Olovi, 74х968 мм, ламинированный, серый',
    'Порог дверной Olovi, 70х937 мм, ламинированный, дуб классик',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.9814, 0.5158],
#         [0.9814, 1.0000, 0.5642],
#         [0.5158, 0.5642, 1.0000]])

Evaluation

Metrics

Triplet

• Dataset: hard-neg-eval
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9962

Training Details

Training Dataset

Unnamed Dataset

• Size: 123,565 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 5 tokens mean: 17.49 tokens max: 44 tokens	min: 7 tokens mean: 21.8 tokens max: 44 tokens	min: 7 tokens mean: 20.83 tokens max: 42 tokens

• Samples:

  anchor	positive	negative
  Заклёпка Matrix 4.8x16 мм 50 шт.	Заклепка вытяжная комбинированная SWFS 4,8х16 (100 шт)	Анкер забивной SWFS ZY М16 (25 шт)
  Ручка дверная на розетке Фабрика замков 016 алюминий гальваническое покрытие цвет матовый хром	Ручка для финских дверей SCHLOSS 01001 006 хром (10/50)	Защелка дверная SCHLOSS 42031 KL-01 с ручкой шар хром (30)
  решетка вентиляционная 220х220 пластик	Решетка вентиляционная 220х220 мм, МД2222, пластик	Решетка вентиляционная вытяжная 1825П, 183х253мм, пластик

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 13,730 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 5 tokens mean: 17.45 tokens max: 47 tokens	min: 6 tokens mean: 21.91 tokens max: 42 tokens	min: 9 tokens mean: 20.6 tokens max: 40 tokens

• Samples:

  anchor	positive	negative
  Линолеум полукоммерческий таркетт Santiago 3м	Линолеум полукоммерческий Polystyl Space F Santiago 1 (3 м), пр-во ТАРКЕТТ	Линолеум полукоммерческий Juteks Master Ester 3 (4 м)
  Штуцер 5-ти выводной 1"x1/4" ВР-НР латунь	Штуцер 5-ти выводной для насосов 1"ВР х 1"ВР х 1"НР х 1/4"ВР х 1/4"НР	Эксцентрик для смесителя 3/4"x1/2" НР
  Вагонка евро хвоя 12.5х88х2500 мм	Вагонка Евро 12,5х88(96)x2500 мм сорт С хвоя (10 шт)	Вагонка Евро 12,5х88(96)x2700 мм сорт С хвоя (10 шт)

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• gradient_accumulation_steps: 8
• learning_rate: 1e-05
• weight_decay: 0.01
• num_train_epochs: 10
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• load_best_model_at_end: True
• optim: adamw_torch
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 8
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 1e-05
• weight_decay: 0.01
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 10
• max_steps: -1
• lr_scheduler_type: cosine
• lr_scheduler_kwargs: None
• 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
• bf16: False
• fp16: False
• 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: True
• 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}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch
• optim_args: None
• adafactor: False
• group_by_length: False
• length_column_name: length
• project: huggingface
• trackio_space_id: trackio
• 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: no
• 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: True
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	hard-neg-eval_cosine_accuracy
0.1036	50	2.2199	-	-
0.2071	100	0.4732	0.0789	0.9821
0.3107	150	0.0994	-	-
0.4143	200	0.0739	0.0485	0.9886
0.5179	250	0.0664	-	-
0.6214	300	0.0567	0.0397	0.9893
0.7250	350	0.0517	-	-
0.8286	400	0.0439	0.0315	0.9913
0.9322	450	0.0445	-	-
1.0352	500	0.0399	0.0278	0.9926
1.1388	550	0.0346	-	-
1.2424	600	0.0343	0.0245	0.9937
1.3459	650	0.0309	-	-
1.4495	700	0.0304	0.0223	0.9942
1.5531	750	0.0287	-	-
1.6567	800	0.0293	0.0203	0.9949
1.7602	850	0.0283	-	-
1.8638	900	0.0269	0.0195	0.9948
1.9674	950	0.0265	-	-
2.0704	1000	0.0235	0.0182	0.9953
2.1740	1050	0.0238	-	-
2.2776	1100	0.0222	0.0170	0.9958
2.3811	1150	0.0189	-	-
2.4847	1200	0.021	0.0168	0.9955
2.5883	1250	0.0215	-	-
2.6919	1300	0.0217	0.0162	0.9954
2.7954	1350	0.021	-	-
2.8990	1400	0.0197	0.0157	0.9963
3.0021	1450	0.0199	-	-
3.1056	1500	0.0165	0.0153	0.9960
3.2092	1550	0.0158	-	-
3.3128	1600	0.0165	0.0148	0.9962

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.12.3
• Sentence Transformers: 5.2.0
• Transformers: 4.57.6
• PyTorch: 2.9.1+cu128
• Accelerate: 1.12.0
• Datasets: 4.5.0
• 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",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}