README.md

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:21123868
  - loss:CachedMultipleNegativesRankingLoss
base_model: sentence-transformers/paraphrase-multilingual-mpnet-base-v2
widget:
  - source_sentence: 系统管理员技术员——TS/SCI级别并拥有多项式验证
    sentences:
      - >-
        support development of annual budget, create a financial report, report
        analysis results, Microsoft Access, accounting, use presentation
        software, interpret financial statements, synthesise financial
        information, develop vaccines, handle financial overviews of the store,
        produce statistical financial records, develop financial statistics
        reports, explain accounting records, financial analysis, SAP R3,
        represent the company, examine budgets, prepare presentation material,
        use spreadsheets software, forecast account metrics, meet deadlines,
        prepare financial projections, manage budgets, exercise self-control,
        financial statements
      - >-
        ensure cross-department cooperation, establish customer rapport, improve
        business processes, manage technical security systems, handle incidents,
        maintain ICT system, explain characteristics of computer peripheral
        equipment, gather technical information, collaborate in company's daily
        operations , apply change management, maintain technical equipment,
        communicate with customers, solve technical problems, perform ICT
        troubleshooting, use ICT equipment in maintenance activities, manage
        major incidents, build business relationships, computer engineering,
        perform software recovery testing, identify process improvements,
        maintain relationship with customers, carry out project activities,
        collaborate in the development of marketing strategies, computer
        technology, technical terminology
      - >-
        utilise machine learning, cloud technologies, develop predictive models,
        assess sportive performance, formulate findings , principles of
        artificial intelligence, perform business research, communicate with
        stakeholders, computer engineering, build predictive models, computer
        science, develop automated software tests, analyse business objectives,
        Agile development, cloud monitoring and reporting, provide written
        content, obtain relevant licenses, design prototypes, machine learning,
        e-learning software infrastructure, analyse education system,
        disseminate results to the scientific community, learning technologies,
        ML (computer programming), task algorithmisation
  - source_sentence: 安全运营官
    sentences:
      - >-
        deliver outstanding service, manage carriers, direct customers to
        merchandise, improve customer interaction, manage time, support
        managers, assist customers, process customer orders, manage customer
        service, satisfy customers, guarantee customer satisfaction, respond to
        customers' inquiries
      - >-
        manage several projects, implement operational business plans, identify
        improvement actions, develop strategy to solve problems, manage website,
        carry out project activities, follow reporting procedures, supervise
        site maintenance, adjust priorities, schedule shifts, conduct public
        presentations, motivate others, manage operational budgets, report to
        the team leader, encourage teams for continuous improvement, lead the
        sustainability reporting process, implement sustainable procurement,
        show an exemplary leading role in an organisation, manage manufacturing
        facilities, develop training programmes, develop production line, supply
        chain management, leadership principles, lead a team, coaching
        techniques
      - >-
        provide emergency supplies, provide first aid, liaise with security
        authorities, apply medical first aid in case of emergency, regulate
        traffic, train security officers, maintain physical fitness, provide
        protective escort, ensure public safety and security, ensure inspections
        of facilities, work in inclement conditions, follow procedures in the
        event of an alarm, set safety and security standards, comply with the
        principles of self-defence, present reports, maintain facility security
        systems, conduct security screenings, types of evaluation , monitor
        security measures, office equipment, escort pedestrians across streets,
        advise on security staff selection, wear appropriate protective gear,
        work in outdoor conditions, assist emergency services
  - source_sentence: Empleado de control de COVID
    sentences:
      - >-
        maintain records of clients' prescriptions, assist people in
        contaminated areas, label samples, maintain museum records, apply social
        distancing protocols, collect biological samples from patients,
        infection control, label medical laboratory samples, disinfect surfaces,
        maintain customer records, ensure health and safety of staff, personal
        protective equipment, remove contaminated materials, store contaminated
        materials, prepare prescription labels, use personal protection
        equipment
      - >-
        promote organisational communication, provide legal advice, human
        resource management, company policies, perform customer management,
        business processes, ensure compliance with legal requirements, develop
        communications strategies, enforce company values, develop outreach
        training plans, use consulting techniques, develop employment policies,
        human resources department processes, personnel management, identify
        training needs, participate in health personnel training, health and
        safety in the workplace, lead police investigations, ensure compliance
        with policies, prepare compliance documents, perform internal
        investigations, develop employee retention programs, develop corporate
        training programmes, customer relationship management, manage
        localisation
      - >-
        perform escalation procedure, imprint visionary aspirations into the
        business management, observe confidentiality, impart business plans to
        collaborators, lead a team, human resources department processes,
        respect confidentiality obligations, hire human resources, manage
        commercial risks, develop business plans, communicate with stakeholders,
        maintain relationship with customers, manage several projects, provide
        improvement strategies, manage technical security systems, knowledge
        management, risk management, develop program ideas, perform project
        management, project management, cope with uncertainty, address
        identified risks, provide performance feedback, information
        confidentiality, track key performance indicators
  - source_sentence: Aerie - Brand Ambassador (Sales Associate) - US
    sentences:
      - >-
        lay bricks, provide first aid, enforce park rules, conflict management,
        give swimming lessons, assist in performing physical exercises, perform
        park safety inspections, assist in the movement of heavy loads, lead a
        team, first aid, supervise pool activities, swim, coach staff for
        running the performance, show an exemplary leading role in an
        organisation, teach public speaking principles, collaborate with
        coaching team, supervise work, calculate stairs rise and run, calculate
        compensation payments, manage a team, information confidentiality
      - >-
        react to events in time-critical environments, operate in a specific
        field of nursing care, clinical science, promote healthy fitness
        environment, lead others, comply with legislation related to health
        care, maintain a safe, hygienic and secure working environment, provide
        healthcare services to patients in specialised medicine, write English,
        conduct physical examinations, leadership principles, use clinical
        assessment techniques, apply context specific clinical competences,
        conduct health related research, conceptualise healthcare user’s needs,
        assessment processes, communicate in healthcare, provide professional
        care in nursing, nursing science, promote health and safety, implement
        policy in healthcare practices, engage with stakeholders, identify
        problems, respond to changing situations in health care, perform
        resource planning
      - >-
        ensure the privacy of guests, provide customised products, company
        policies, exude enthusiasm during the action sessions, provide customer
        guidance on product selection, collect briefing regarding products,
        perform multiple tasks at the same time, create solutions to problems,
        respond to visitor complaints
  - source_sentence: 医师——危重症护理——重症监护专家——项目医务总监
    sentences:
      - >-
        handle incidents, provide technical documentation, coordinate
        operational activities, ensure information security, work in teams,
        manage manufacturing documentation, project configuration management,
        operate call distribution system, maintain computer hardware, apply
        change management, manage aircraft support systems, perform escalation
        procedure, manage production changeovers, maintenance operations,
        call-centre technologies, manage service contracts in the drilling
        industry, encourage teambuilding, manage major incidents, resolve
        equipment malfunctions, work independently, think analytically, manage
        maintenance operations, maintain plan for continuity of operations
      - >-
        develop recycling programs, receive actors' resumes, work in cold
        environments, perform cleaning duties, operate floor cleaning equipment,
        operate forklift
      - >-
        perform technical tasks with great care, supervise medical residents,
        manage a multidisciplinary team involved in patient care, administrative
        tasks in a medical environment, demonstrate technical skills during
        neurological surgery, apply problem solving in social service, intensive
        care medicine, provide comprehensive care for patients with surgical
        conditions, work in teams, solve problems
pipeline_tag: sentence-similarity
library_name: sentence-transformers
co2_eq_emissions:
  emissions: 717.3535184611766
  energy_consumed: 1.9440474755045436
  source: codecarbon
  training_type: fine-tuning
  on_cloud: true
  cpu_model: Intel(R) Xeon(R) CPU @ 2.20GHz
  ram_total_size: 83.47684860229492
  hours_used: 5.34
  hardware_used: 1 x NVIDIA A100-SXM4-40GB
license: mit
language:
  - en
  - es
  - de
  - zh
  - mul
  - multilingual

SentenceTransformer based on sentence-transformers/paraphrase-multilingual-mpnet-base-v2

This is a sentence-transformers model specifically trained for job title matching and similarity. It's finetuned from sentence-transformers/paraphrase-multilingual-mpnet-base-v2 on a large dataset of job titles and their associated skills/requirements across multiple languages. The model maps English, Spanish, German and Chinese job titles and descriptions to a 1024-dimensional dense vector space and can be used for semantic job title matching, job similarity search, and related HR/recruitment tasks.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: sentence-transformers/paraphrase-multilingual-mpnet-base-v2
• Maximum Sequence Length: 64 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset: 4 x 5.2M high-quality job title - skills pairs in English, Spanish, German and Chinese

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': 64, 'do_lower_case': False}) with Transformer model: XLMRobertaModel 
  (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): Asym(
    (anchor-0): Dense({'in_features': 768, 'out_features': 1024, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
    (positive-0): Dense({'in_features': 768, 'out_features': 1024, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
  )
)

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load and use the model with the following code:

import torch
import numpy as np
from tqdm.auto import tqdm
from sentence_transformers import SentenceTransformer
from sentence_transformers.util import batch_to_device, cos_sim

# Load the model
model = SentenceTransformer("TechWolf/JobBERT-v3")

def encode_batch(jobbert_model, texts):
    features = jobbert_model.tokenize(texts)
    features = batch_to_device(features, jobbert_model.device)
    features["text_keys"] = ["anchor"]
    with torch.no_grad():
        out_features = jobbert_model.forward(features)
    return out_features["sentence_embedding"].cpu().numpy()

def encode(jobbert_model, texts, batch_size: int = 8):
    # Sort texts by length and keep track of original indices
    sorted_indices = np.argsort([len(text) for text in texts])
    sorted_texts = [texts[i] for i in sorted_indices]
    
    embeddings = []
    
    # Encode in batches
    for i in tqdm(range(0, len(sorted_texts), batch_size)):
        batch = sorted_texts[i:i+batch_size]
        embeddings.append(encode_batch(jobbert_model, batch))
    
    # Concatenate embeddings and reorder to original indices
    sorted_embeddings = np.concatenate(embeddings)
    original_order = np.argsort(sorted_indices)
    return sorted_embeddings[original_order]

# Example usage
job_titles = [
    'Software Engineer',
    '高级软件开发人员',  # senior software developer
    'Produktmanager',  # product manager
    'Científica de datos'  # data scientist
]

# Get embeddings
embeddings = encode(model, job_titles)

# Calculate cosine similarity matrix
similarities = cos_sim(embeddings, embeddings)
print(similarities)

The output will be a similarity matrix where each value represents the cosine similarity between two job titles:

tensor([[1.0000, 0.8087, 0.4673, 0.5669],
        [0.8087, 1.0000, 0.4428, 0.4968],
        [0.4673, 0.4428, 1.0000, 0.4292],
        [0.5669, 0.4968, 0.4292, 1.0000]])

Training Details

Training Dataset

Unnamed Dataset

• Size: 21,123,868 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 4 tokens mean: 10.56 tokens max: 38 tokens	min: 19 tokens mean: 61.08 tokens max: 64 tokens

• Samples:

  anchor	positive
  通信与培训专员	deliver online training, liaise with educational support staff, interact with an audience, construct individual learning plans, lead a team, develop corporate training programmes, learning technologies, communication, identify with the company's goals, address an audience, learning management systems, use presentation software, motivate others, provide learning support, engage with stakeholders, identify skills gaps, meet expectations of target audience, develop training programmes
  Associate Infrastructure Engineer	create solutions to problems, design user interface, cloud technologies, use databases, automate cloud tasks, keep up-to-date to computer trends, work in teams, use object-oriented programming, keep updated on innovations in various business fields, design principles, Angular, adapt to changing situations, JavaScript, Agile development, manage stable, Swift (computer programming), keep up-to-date to design industry trends, monitor technology trends, web programming, provide mentorship, advise on efficiency improvements, adapt to change, JavaScript Framework, database management systems, stimulate creative processes
  客户顾问/出纳	customer service, handle financial transactions, adapt to changing situations, have computer literacy, manage cash desk, attend to detail, provide customer guidance on product selection, perform multiple tasks at the same time, carry out financial transactions, provide membership service, manage accounts, adapt to change, identify customer's needs, solve problems

• Loss: CachedMultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "mini_batch_size": 512
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• overwrite_output_dir: True
• per_device_train_batch_size: 2048
• per_device_eval_batch_size: 2048
• num_train_epochs: 1
• fp16: True

All Hyperparameters

Click to expand

• overwrite_output_dir: True
• do_predict: False
• eval_strategy: no
• prediction_loss_only: True
• per_device_train_batch_size: 2048
• per_device_eval_batch_size: 2048
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 5e-05
• weight_decay: 0.0
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1.0
• num_train_epochs: 1
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.0
• 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: False
• fp16: True
• 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: False
• 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
• 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
• 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
• dispatch_batches: None
• split_batches: 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
• eval_use_gather_object: False
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss
0.0485	500	3.89
0.0969	1000	3.373
0.1454	1500	3.1715
0.1939	2000	3.0414
0.2424	2500	2.9462
0.2908	3000	2.8691
0.3393	3500	2.8048
0.3878	4000	2.7501
0.4363	4500	2.7026
0.4847	5000	2.6601
0.5332	5500	2.6247
0.5817	6000	2.5951
0.6302	6500	2.5692
0.6786	7000	2.5447
0.7271	7500	2.5221
0.7756	8000	2.5026
0.8240	8500	2.4912
0.8725	9000	2.4732
0.9210	9500	2.4608
0.9695	10000	2.4548

Environmental Impact

Carbon emissions were measured using CodeCarbon.

• Energy Consumed: 1.944 kWh
• Carbon Emitted: 0.717 kg of CO2
• Hours Used: 5.34 hours

Training Hardware

• On Cloud: Yes
• GPU Model: 1 x NVIDIA A100-SXM4-40GB
• CPU Model: Intel(R) Xeon(R) CPU @ 2.20GHz
• RAM Size: 83.48 GB

Framework Versions

• Python: 3.10.16
• Sentence Transformers: 4.1.0
• Transformers: 4.48.3
• PyTorch: 2.6.0+cu126
• Accelerate: 1.3.0
• Datasets: 3.5.1
• Tokenizers: 0.21.0

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",
}

CachedMultipleNegativesRankingLoss

@misc{gao2021scaling,
    title={Scaling Deep Contrastive Learning Batch Size under Memory Limited Setup},
    author={Luyu Gao and Yunyi Zhang and Jiawei Han and Jamie Callan},
    year={2021},
    eprint={2101.06983},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}