dnth/ssf-retriever-modernbert-embed-base-v4.2

SentenceTransformer based on nomic-ai/modernbert-embed-base

This is a sentence-transformers model finetuned from nomic-ai/modernbert-embed-base on the ssf-train-valid-v4.2 dataset. 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: nomic-ai/modernbert-embed-base
• Maximum Sequence Length: 8192 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • ssf-train-valid-v4.2

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': 8192, '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})
  (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("dnth/ssf-retriever-modernbert-embed-base-v4.2")
# Run inference
sentences = [
    'The Senior Manufacturing Planning Executive formulates production plans and organises materials, manpower and resources to accomplish manufacturing functions to fulfil customer and financial commitments. He/She validates the master production schedule (MPS) and drives adherence of manufacturing works to project schedules and goals in collaboration with cross-functional leads. He leads material requirements planning and programme reviews with relevant stakeholders. He is responsible for optimising supply chain and logistics planning, contract negotiations, vendor sourcing, inventory planning and control and warehousing operations to meet manufacturing requirements. He leverages data from supply chain management (SCM) systems to enhance decision-making and implements supplier capability development plans to enhance performance. He drives continuous improvements on product on-time delivery and total available man-hours, develops strategies and priorities for critical customer issues, facilitates problem-solving, leads in regular reviews with customers and suppliers, and establishes best practices on process improvements to enhance productivity. He proactively contributes to the development of lean and sustainability practices, and conducts research and digital innovation in targeted areas for continuous process improvements. As a team leader, he appraises staff performance and conducts coaching and mentoring for planning personnel. He should possess advanced statistical, forecasting and analytical skills to predict planning and resource requirements. He is able to drive cross-functional collaboration between internal and external stakeholders to optimise the planning processes and ensure maximum resource utilisation.',
    'The Senior Manufacturing Planning Executive develops and implements production schedules while coordinating materials, workforce, and resources to meet manufacturing targets aligned with customer demands and financial objectives. This role involves validating the master production schedule and ensuring manufacturing activities comply with project timelines through collaboration with various departments. The executive leads material planning and program assessments with key partners and is accountable for optimizing supply chain logistics, managing contracts, sourcing vendors, controlling inventory, and overseeing warehouse operations to support manufacturing needs. By utilizing supply chain management data, the executive enhances decision-making and drives supplier capability improvements. They champion continuous improvements in on-time delivery performance and labor efficiency, formulate strategies to address critical customer concerns, facilitate problem resolution, conduct stakeholder reviews, and promote best practices to boost productivity. Additionally, the role supports lean methodologies and sustainability initiatives, explores digital innovations, and leads process enhancements. As a leader, the executive evaluates team performance and provides coaching and mentoring to planning staff. The position demands strong statistical, forecasting, and analytical expertise to anticipate planning and resource demands and fosters effective collaboration among internal and external partners to maximize planning efficiencies.',
    'The Senior Procurement Executive manages the acquisition of goods and services, negotiates supplier contracts, and oversees vendor relationships to support the company’s purchasing needs. This role focuses on sourcing strategies, supplier evaluations, cost analysis, and procurement compliance within the manufacturing industry. The executive leads procurement planning, coordinates with finance and operations teams, and ensures timely delivery of purchased materials. They are responsible for maintaining supplier performance metrics, conducting market research, and implementing procurement best practices. The role requires strong negotiation skills, supplier risk management, and contract administration experience. As a senior professional, the executive supervises procurement staff and drives continuous improvement initiatives in sourcing processes.',
]
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, 0.9240, 0.5197],
#         [0.9240, 1.0000, 0.5085],
#         [0.5197, 0.5085, 1.0000]])

Training Details

Training Dataset

ssf-train-valid-v4.2

• Dataset: ssf-train-valid-v4.2 at 97c8b4d
• Size: 7,540 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 58 tokens mean: 167.85 tokens max: 355 tokens	min: 58 tokens mean: 138.3 tokens max: 293 tokens	min: 50 tokens mean: 108.71 tokens max: 249 tokens

• Samples:

  anchor	positive	negative
  The Manufacturing Engineer/Production Engineer (Assembly) develops detailed operation and specification sheets throughout the assembly cycle. He/She coordinates shop floor operations and process control, and plans resources to meet production targets. He is conversant with tools and fixtures design and computer integrated manufacturing (CIM) technologies. He determines appropriate resources and processes for engineering application while ensuring working conditions of assembly equipment and machinery. He also manages assembly techniques and verifies conformance of new aircraft components and parts to specifications. He ensures adherence of assembly operations to legislative and airworthiness requirements, as well as with the organisation's standard operating procedures (SOPs), safety, health and quality systems. He identifies opportunities for continuous improvement through data analytics, research and innovation, and implements lean and sustainability practices in assembly. He monitor...	The Manufacturing Engineer (Assembly) is responsible for creating detailed operation and specification documentation for the assembly process. This role involves coordinating shop floor activities and overseeing process controls while managing resource planning to achieve production goals. The engineer applies expertise in tooling and fixture design alongside computer integrated manufacturing (CIM) technologies to determine suitable resources and processes for engineering tasks. Ensuring optimal working conditions for assembly machinery and equipment, the engineer supervises assembly methods and confirms that new aircraft parts meet stringent specification requirements. Compliance with legislative, airworthiness, and organizational standard operating procedures (SOPs), as well as health, safety, and quality management systems, is rigorously maintained. The role focuses on identifying and implementing continuous improvement initiatives through data analysis, innovation, and lean manufac...	The Manufacturing Engineer (Quality Assurance) oversees the inspection and testing of finished products to ensure compliance with quality standards. This role manages quality control procedures throughout the production cycle but does not directly engage in assembly or shop floor coordination. The engineer utilises statistical process control and quality management systems to monitor product conformity, focusing on defect reduction and customer satisfaction. Responsibilities include conducting audits, documenting non-conformance issues, and recommending corrective actions aligned with regulatory requirements and internal policies. While familiar with manufacturing tools and technologies, this position emphasizes quality assurance processes rather than resource planning or tooling design. The engineer collaborates with cross-functional teams to improve product reliability and supports training initiatives on quality standards. Strong analytical skills and attention to detail are necessa...
  The Linen Room Attendant/Laundry Valet Attendant performs daily assigned duties to support the day-to-day laundry, linen and uniform room operations, ensuring the delivery of clean garments, uniforms, towels and linens to all internal and external customers. He/She collects and delivers guest laundry, performs laundry cleaning, sorts and issues linens and uniforms, and assists in inventory count. He also cleans and maintains laundry equipment and the work area. As part of service delivery, the Linen Room Attendant/Laundry Valet Attendant has to handle guests' requests and respond to their concerns and feedback in a professional and courteous manner. He complies with organisational guidelines and regulations on hygiene and workplace safety and health, and reports safety hazards observed to ensure workplace safety and security. He is a team player with a high level of attentiveness to details and good communication skills to interact with guests and all levels of staff. He works on shift...	The Linen Room Attendant/Laundry Valet Attendant is responsible for supporting daily operations in the laundry, linen, and uniform rooms by ensuring prompt and efficient delivery of cleaned garments, towels, uniforms, and linens to both internal departments and external guests. This role involves collecting and returning guest laundry, sorting and distributing linens and uniforms, conducting inventory checks, and maintaining cleanliness of laundry equipment and workspaces. The attendant addresses guest inquiries and concerns professionally and courteously, adheres to hygiene and workplace safety standards, and promptly reports any safety issues. This position requires teamwork, attention to detail, effective communication skills, and physical stamina to handle tasks such as standing, walking, and lifting heavy laundry loads throughout shifts that may include weekends and public holidays.	The Linen Room Supervisor oversees the strategic planning and management of laundry services within a hotel, leading a team of attendants and coordinating with multiple departments to optimize operational efficiency. This senior role involves budgeting, staff training, and implementing quality control measures rather than performing hands-on laundry tasks. The supervisor is responsible for developing service standards, managing vendor relationships, and ensuring compliance with corporate policies, with minimal direct involvement in daily linen sorting or equipment maintenance. Strong leadership, decision-making capabilities, and experience in workforce management are essential, while physical demands are limited compared to frontline laundry roles.
  The General Worker / Operator performs general duties, and cleaning and housekeeping tasks as assigned. He/She is required to assist in operating machinery under supervision and moving aircraft components, equipment and materials from the store to respective work areas. He is expected to adhere to the organisation's standard operating procedures (SOPs), and safety, health and quality systems. He supports in implementation of continuous improvement initiatives to ensure workspace efficiency and effectiveness. He works in a hangar or workshop and may be required to work in shifts. He should be comfortable with repetitive work activities and exposure to physically demanding work conditions such as long standing hours and extreme temperatures.	The General Worker / Operator is responsible for carrying out various general tasks including cleaning and housekeeping duties as directed. This role involves assisting with machinery operation under guidance and transporting aircraft parts, equipment, and supplies from storage to designated work areas. The incumbent must strictly follow the company’s standard operating procedures, along with safety, health, and quality protocols. They contribute to continuous improvement efforts aimed at enhancing workspace productivity and efficiency. The position is based in a hangar or workshop environment and may require shift work. The ideal candidate should be able to handle repetitive tasks and endure physically challenging conditions such as prolonged standing and exposure to temperature extremes.	The Warehouse Clerk manages inventory records and coordinates the receipt and dispatch of goods within the logistics sector. This role requires proficiency in inventory management software and strong organizational skills to maintain stock accuracy. The Warehouse Clerk operates in a distribution center and collaborates closely with supply chain teams to ensure timely delivery schedules. The position demands attention to detail and the ability to work under pressure but does not involve machinery operation or physically strenuous activities common in manufacturing environments.

• Loss: MultipleNegativesRankingLoss with these parameters:

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

Evaluation Dataset

ssf-train-valid-v4.2

• Dataset: ssf-train-valid-v4.2 at 97c8b4d
• Size: 1,885 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 60 tokens mean: 170.26 tokens max: 403 tokens	min: 59 tokens mean: 138.72 tokens max: 265 tokens	min: 50 tokens mean: 109.98 tokens max: 252 tokens

• Samples:

  anchor	positive	negative
  The Technician (Signal and Communications) works in a team to perform preventive and corrective maintenance of signal, communication and control systems, to improve the reliability of signal, communication and control systems. He/She assists in the preparation of maintenance activities and is technically inclined and adept in handling electronics and computer-based systems and equipment for maintenance. He also supervises the work of contractors and external stakeholders in ensuring adherence to operating requirements and safety standards. He may be required to perform shift duties at various rail premises such as workshops, depots, train stations, and train tunnels. He is capable of communicating effectively within the team, is able to multi-task and can prioritises his assigned maintenance workload in supporting maintenance activities.	The Technician (Signal and Communications) collaborates within a team to conduct routine and emergency maintenance on signal, communication, and control infrastructures, aiming to enhance system reliability. This role involves assisting in the planning of maintenance operations and requires strong technical skills in electronics and computer-based maintenance tools. The technician oversees contractors and external partners to ensure compliance with operational protocols and safety guidelines. Shift work at various rail facilities, including workshops, depots, stations, and tunnels, may be necessary. Effective team communication, multitasking abilities, and prioritization of maintenance tasks are essential to support ongoing maintenance efforts.	The Technician (Electrical Installations) is responsible for installing and testing electrical wiring and equipment in residential and commercial buildings. They prepare site layouts, follow electrical codes, and ensure safety during installation processes. The technician coordinates with suppliers and clients but does not engage in signal or communication system maintenance. Shift work is generally not required, and the role focuses on hands-on installation rather than supervising external contractors. Strong knowledge of electrical wiring, circuit breakers, and household electrical standards is necessary, along with good communication skills to liaise with homeowners and site managers.
  The Visual Merchandiser manages shopper marketing activities and is responsible for the conceptualisation of the visual merchandising plans. He/she oversees the set-up of merchandise display by coaching in-store teams. He is also responsible for market research efforts relating to visual merchandising. He operates in a fast-paced and creative environment where he conceptualises eye-catching product displays, store layouts and designs to promote the store's products. He is creative, detail-oriented and is effective working within tight deadlines. He is able to effectively prioritise multiple assignments and possesses an aesthetic flair.	The Visual Merchandiser is responsible for planning and executing shopper marketing strategies through innovative visual displays. This role involves guiding retail teams in arranging merchandise presentations and ensuring the store environment is appealing and aligned with brand standards. The Visual Merchandiser conducts market research to stay updated on trends and consumer preferences, working in a dynamic, fast-paced setting that demands creativity and precision. Strong organizational skills and an eye for design are essential to manage multiple projects and deliver compelling store layouts that enhance customer engagement.	The Visual Merchandiser leads the digital marketing campaigns for retail brands, focusing on online shopper engagement and social media promotions. He/she develops content strategies, coordinates with creative teams, and analyses ecommerce data to optimise product visibility. Operating in a technology-driven environment, the Visual Merchandiser applies analytical skills and marketing knowledge to influence buying behaviour through digital channels rather than physical displays. This role requires proficiency in digital tools and a strong understanding of consumer analytics rather than traditional visual merchandising techniques.
  The Network Development Technician implements gas transmission and/or distribution network development projects and monitors site activities. He/She supports the preparation of construction activity records, project progress reports and materials required for payments. He also liaises with contractors and customers to carry out metering works and performs the installation, testing and commissioning of residential meters. He applies Safe System of Work (SSoW) procedures and risk control measures to ensure work activities are carried out safely, and in compliance with Workplace Safety and Health (WSH) Act. He is a member of the Emergency Response Team and follows emergency response plans and relevant safety procedures. He occasionally works at construction sites for the gas transmission and/or distribution network development projects. He is a good team player who collaborates and communicates effectively with key stakeholders. He is detailed in ensuring that operations are carried out a...	The Network Development Technician is responsible for executing gas transmission and distribution network expansion initiatives while overseeing on-site operations. This role involves assisting in the documentation of construction activities, compiling project status updates, and coordinating materials for billing purposes. The technician interacts with contractors and clients to facilitate metering installations, including the testing and commissioning of residential gas meters. Adherence to Safe System of Work protocols and risk mitigation strategies is essential to maintain compliance with the Workplace Safety and Health Act. As an integral member of the Emergency Response Team, the technician follows prescribed emergency procedures and safety guidelines. Fieldwork at construction locations is periodically required. Strong teamwork, clear communication with stakeholders, and meticulous attention to procedural compliance are key attributes for success in this role.	The Network Operations Coordinator oversees the scheduling and administration of telecommunications network services, ensuring seamless connectivity and customer satisfaction. This position requires coordinating with service providers and vendors to manage infrastructure upgrades and maintenance tasks. The coordinator prepares operational reports and assists with billing reconciliations. Familiarity with IT systems and network management software is essential, alongside strong communication skills to liaise with internal teams and external partners. While safety protocols are observed, the role primarily focuses on service delivery rather than physical installation or emergency response activities. The coordinator works mainly in an office environment and supports multiple projects simultaneously without direct involvement in gas transmission or distribution networks.

• Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 32
• gradient_accumulation_steps: 32
• learning_rate: 2e-05
• weight_decay: 0.01
• num_train_epochs: 5
• lr_scheduler_type: cosine
• warmup_ratio: 0.1
• bf16: True
• tf32: True
• load_best_model_at_end: True
• gradient_checkpointing: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 32
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 32
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 2e-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: 5
• max_steps: -1
• lr_scheduler_type: cosine
• 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: True
• fp16: False
• fp16_opt_level: O1
• half_precision_backend: auto
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: True
• 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}
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• 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: None
• hub_always_push: False
• hub_revision: None
• gradient_checkpointing: True
• 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: False
• 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: False
• prompts: None
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.3390	5	0.163	-
0.6780	10	0.0257	-
1.0	15	0.0048	0.0057
1.3390	20	0.0031	-
1.6780	25	0.0021	-
2.0	30	0.0015	0.0027
2.3390	35	0.0021	-
2.6780	40	0.0023	-
3.0	45	0.001	0.0017
3.3390	50	0.0013	-
3.6780	55	0.0014	-
4.0	60	0.0013	0.0015
4.3390	65	0.0013	-
4.6780	70	0.001	-
5.0	75	0.0018	0.0015

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.12.8
• Sentence Transformers: 5.1.0
• Transformers: 4.55.0
• PyTorch: 2.8.0+cu128
• Accelerate: 1.10.0
• Datasets: 4.0.0
• Tokenizers: 0.21.4

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}
}