SentenceTransformer based on mixedbread-ai/mxbai-embed-large-v1

This is a sentence-transformers model finetuned from mixedbread-ai/mxbai-embed-large-v1. It maps sentences & paragraphs to a 1024-dimensional dense vector space and can be used for retrieval.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: mixedbread-ai/mxbai-embed-large-v1
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 1024 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': 1024, 'pooling_mode': 'cls', '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("sentence_transformers_model_id")
# Run inference
queries = [
    'PETROCHEMICAL INDUSTRY HAZARDOUS & HIGH-INTEGRITY PROCESSES Oswar Rotocorp delivers engineered flow control and material',
]
documents = [
    'PETROCHEMICAL INDUSTRY HAZARDOUS & HIGH-INTEGRITY PROCESSES Oswar Rotocorp delivers engineered flow control and material handling solutions designed specifically for safety- sensitive petrochemical applications. Our equipment supports controlled feeding, pressure containment, and sealed material transfer in refineries, polymer plants, and chemical processing units. Each solution is developed to ensure operational reliability, structural strength, and long-term durability in demanding petrochemical environments. CRITICAL PROCESS AREAS CATALYST POLYMER & RESIN SILO & HOPPER DUST & VAPOR HANDLING & POWDER ISOLATION CONTAINMENT INJECTION DISCHARGE SYSTEMS SYSTEMS BUILT FOR PRESSURE, CONTAINMENT & DURABILITY Petrochemical applications demand precision engineering and structural robustness. Our equipment is manufactured using high-grade materials, reinforced housings, and optimized sealing systems to ensure containment and resistance to corrosive or reactive substances. Rotary Valves are engineered with controlled tolerances and advanced shaft sealing configurations to maintain airlock integrity in pressure or vacuum conveying systems. Specialized sealing arrangements reduce leakage risk and enhance process stability. ENGINEERING FEATURES Heavy-Duty Rotary Valves for Polymer & Catalyst Handling Pneumatic Pressure Rotary Valves for Pressurized Conveying',
    '4.4 Alle in de Overeenkomst vermelde prijzen zijn in euro’s, exclusief btw. Algemene inkoopvoorwaarden – May 2022 – version 1.0 2 5. Levering 5.1 De in de Overeenkomst vermelde uitvoeringstermijn of leveringsdatum is voor Leverancier een fatale termijn respectievelijk fatale datum, tenzij Partijen schriftelijk anders zijn overeengekomen. Het enkel overschrijden van die termijn c.q. datum brengt Leverancier in verzuim, zonder dat een ingebrekestelling is vereist. DMN-WESTINGHOUSE heeft het recht om eenzijdig de leveringstermijn dan wel -datum te verlengen c.q. uit te stellen. 5.2 Indien geen uitdrukkelijke uitvoeringstermijn of -datum is overeengekomen, geldt tussen Partijen een redelijke uitvoerings-/leveringstermijn. 5.3 Indien levering in fasen is overeengekomen en Leverancier in gebreke is gebleven met de levering ter zake van een fase, is artikel 5.1 van toepassing en heeft DMN-WESTINGHOUSE het recht toekomstige leveringen te annuleren, zonder daarbij gehouden te zijn tot enige vergoeding van schade aan Leverancier. 5.4 Zodra Leverancier weet of behoort te weten dat de levering niet, niet tijdig of niet naar behoren zal plaatsvinden, geeft hij hiervan onmiddellijk schriftelijk bericht aan DMN-WESTINGHOUSE onder vermelding van de omstandigheden die de aanleiding van deze tekortkoming vormen, alsmede van de datum waarop Leverancier redelijkerwijs verwacht in staat',
    'Precision CNC-machined heavy-duty housing Optimized rotor geometry for stable discharge Enhanced sealing configuration for minimal leakage Reinforced structural construction Advanced wear-resistant surface treatments WWW.OSWARROTOCORP.COM 2 PREMIUM PERFORMANCE ACROSS INDUSTRIES Typical Applications: High-Pressure Pneumatic Conveying Kiln & Clinker Discharge Boiler Ash Handling Mineral & Ore Transfer Large-Capacity Silo Discharge Dense Phase Conveying Systems Industries Served: Cement Manufacturing Power Generation Mining & Mineral Processing Steel & Metallurgical Industries Chemical Processing Plants WWW.OSWARROTOCORP.COM 3 Contact us Copyright © 2026 OSWAR Corporate Office ROTOCORP. All Rights Reserved. Jaipur, Rajasthan India OSWAR® is a registered trademark of Nagpur, Maharashtra India OSWAR ROTOCORP. This brochure makes no offers, representations or Sales Inquiries warranties (express or implied), and info@oswarrotocorp.com information and data contained in this brochure are for general Helpline reference only and may change at any +44 7917 128505 | +91 77422 36474 time.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 1024] [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[0.7314, 0.0433, 0.3386]], dtype=torch.float16)

Training Details

Training Dataset

Unnamed Dataset

• Size: 851 training samples
• Columns: query and passage
• Approximate statistics based on the first 100 samples:

  	query	passage
  type	string	string
  modality	text	text
  details	min: 7 tokensmean: 17.91 tokensmax: 45 tokens	min: 128 tokensmean: 229.17 tokensmax: 256 tokens

• Samples:

  query	passage
  BACKGROUND Established in 1950, DMN-WESTINGHOUSE has been a worldwide, trusted supplier of rotary valves	BACKGROUND Established in 1950, DMN-WESTINGHOUSE has been a worldwide, trusted supplier of rotary valves and diverter valves for decades, serving a large range of dry bulk solids processing industries. True to our promise to provide future-proof value, our experts continue monitoring the performance of our products in practice, including in the context of customer feedback and emerging new technologies and practices. Our staff have developed a complete range of premium diverter valves for the transport of any type of dry bulk solid in powder, granule, or pellet form. We offer plug, tube, flap, and ball diverter valves – all with the DMN-WESTINGHOUSE quality mark. These durable diverters can be further customised to your specifications and are easy to integrate with your current rotary valves. Tell us more about your setup and applications, and we will offer you the perfect diverter valve. MEET THE 2-TDV TUBE DIVERTER VALVE The 2-port Tube Diverter Valve (2-TDV) is a hygienic, USDA-acce...
  What information is in AL AXL AML Dairy MZC I Chaindrive2 MB PDF?	e 155 Sluitring Washer Unterlegscheibe Rondelle 543 Moer Nut Mutter Écrou 160 Verzonken zeskantbout Countersunk bolt Senkkopfschraube Vis à tête fraisée 544 Sluitring Washer Unterlegscheibe Rondelle 161 Tapbout Bolt Kopfschraube Vis à tête 546 Cilinderkop schroef Cilinder bolt Zylinderbolz Vis de fixation 182 Afdichtring Joint ring Dichtungring Bagues de joint 547 Moer Nut Mutter Écrou 548 Sluitring Washer Unterlegscheibe Rondelle 185 Verl.zeskantbouthouder Hexacan bolt holder Verl. bolzen halter Buttoir de vis a tet 600 Indicatiebout Bolt Kopfschraube Vis à tête 201 Rotor Rotor Rotor Rotor 203 Aandrijfas Drive shaft Antriebswelle Arbre de commande 610 Sensor Sensor Sensor Capteur 226 Inlegspie Key Passfeder Clavette 611 Moer Nut Mutter Écrou 228 Veiligheidsaanslag schroef Safety stop screw Sicherheitsanschlagschraube Vis d'arrêt de sécurité 700 Nokkenschijf Cam disc Nockenscheibe Disque à cames 231 Moer Nut Mutter Écrou 235 Borgmoer Locknut Sicherungsmutter Écrou de sûreté 701 Vulblok...
  What information is in 6 3. OSWAR Japan Partnership Overview.pdf?	OSWAR ROTOCORP OSWAR JAPAN PARTNERSHIP A STRATEGIC GLOBAL ALLIANCE Oswar Rotocorp operates under the strategic framework of OSWAR Japan — a powerful industrial collaboration that brings together structured Japanese engineering discipline and advanced Indian manufacturing excellence. This partnership represents a long-term alignment focused on quality consistency, precision standards, and sustainable industrial growth. By combining the technical rigor and systematic approach of Japan with scalable and efficient production capabilities in India, Oswar Rotocorp delivers globally competitive rotary airlock and feeding solutions. KEY FEATURES JAPANESE STRUCTURED GLOBAL SCALABLE INDIAN ENGINEERING QUALITY STANDARD MANUFACTURING DISCIPLINE CONTROL DESIGN STRENGTH SYSTEMS PHILOSOPHY PRECISION ENGINEERED UNDER GLOBAL STANDARDS Under the OSWAR Japan framework, every rotary airlock system is developed through disciplined design methodologies, strict dimensional control, and performance-driven eng...

• 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

• per_device_train_batch_size: 16
• num_train_epochs: 1
• learning_rate: 5e-09
• warmup_steps: 6
• dataloader_drop_last: True
• dataloader_pin_memory: False

All Hyperparameters

• per_device_train_batch_size: 16
• num_train_epochs: 1
• max_steps: -1
• learning_rate: 5e-09
• lr_scheduler_type: linear
• lr_scheduler_kwargs: None
• warmup_steps: 6
• 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: False
• 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: 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: True
• dataloader_num_workers: 0
• dataloader_pin_memory: False
• 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_static_graph: None
• 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: None
• local_rank: -1
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Time

• Training: 4.3 minutes

Framework Versions

• Python: 3.14.4
• Sentence Transformers: 5.5.1
• Transformers: 5.9.0
• PyTorch: 2.12.0
• Accelerate: 1.14.0
• Datasets: 5.0.0
• 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{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},
}