jina-embeddings-v2-small-en_linkedin_profile_model_run1__modified_config

SentenceTransformer based on jinaai/jina-embeddings-v2-small-en

This is a sentence-transformers model finetuned from jinaai/jina-embeddings-v2-small-en. It maps sentences & paragraphs to a 512-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

https://wandb.ai/deklanw/sentence-transformers/runs/frjg0h13?nw=nwuserdeklanw

Model Description

• Model Type: Sentence Transformer
• Base model: jinaai/jina-embeddings-v2-small-en <!-- at revision 44e7d1d6caec8c883c2d4b207588504d519788d0 -->
• Maximum Sequence Length: 2048 tokens
• Output Dimensionality: 512 dimensions
• Similarity Function: Cosine Similarity

<!-- - **Training Dataset:** Unknown --> <!-- - **Language:** Unknown --> <!-- - **License:** Unknown -->

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': 2048, 'do_lower_case': False}) with Transformer model: JinaBertModel 
  (1): Pooling({'word_embedding_dimension': 512, '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})
)

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("borgcollectivegmbh/jina-embeddings-v2-small-en_linkedin_profile_model_run1")
# Run inference
sentences = [
    'Experienced in SAP consultancy',
    '=== PERSON ===\nMärt Ehrenpreis\nN/A\nTallinn Metropolitan Area\n470\n\n=== ROLES ===\n>>> COMPANY ROLES <<<\n\nCompany Role #1:\n  Board Member CTO\n  2023-04 - N/A\n  N/A\n  Elron / Eesti Liinirongid AS \n  eestiliinirongid\n  Truck Transportation\n  Loome Sulle aega – viime Sind kohale kiiresti, turvaliselt ja mugavalt\n  414\n\nCompany Role #2:\n  Managing Director, Proxion Plan Estonia OÜ\n  2021-01 - 2023-04\n  N/A\n  Proxion\n  proxion-plan-oy\n  Civil Engineering\n  Proxion on osa WSP-yhtiöitä.\nOtathan seurantaan WSP in Finland -tilin!\n  2563\n\n>>> SCHOOL ROLES <<<\nNo school roles listed.',
    "=== PERSON ===\nHarvindar Singh Garcha\nHarvindar's passion for technology can be traced back right from his childhood, Getting into the field of computer science bought him to have more focus on his goal. He is a fast learner and easily adapts to new skills. He always has the hunger and curiosity to learn new skills. <br><br>He has expertise in the design &  development of RESTful APIs and back-end services utilizing Python ecosystem, Django, Flask, Docker, and SQL with an emphasis on scalability and security.<br><br>He is currently working as a full stack developer at AI Automotive company where he has worked on technologies Django, Flask, Python, RESTful APIs, React, Redux, Celery, and JavaScript, while keeping in mind all the secure code practices of OWASP top 10.<br><br>His specialties include quickly learning new skills, Programming languages, and Critical thinking in problem-solving.<br><br>In his free time, you will find him reading articles about the latest trending technologies on Internet,  Spending some time on Quora & hiking on the mountains.\nPune, Maharashtra, India\n1294\n\n=== ROLES ===\n>>> COMPANY ROLES <<<\n\nCompany Role #1:\n  Software Development Analyst\n  2020-04 - 2020-08\n  N/A\n  Metta Social\n  metta-social\n  IT Services and IT Consulting\n  Building world’s largest common good platform to enable sustainable impact at scale!\n  4002\n\nCompany Role #2:\n  Junior Backend Engineer\n  2020-09 - 2021-03\n  N/A\n  SRV Media\n  srv-media\n  Advertising Services\n  Insights | Ideas | Impact\n  49579\n\nCompany Role #3:\n  Full Stack Developer\n  2021-03 - 2022-01\n  N/A\n  SRV Media\n  srv-media\n  Advertising Services\n  Insights | Ideas | Impact\n  49579\n\nCompany Role #4:\n  Software Development Intern\n  2019-10 - 2020-03\n  Currently building a B2B social ecosystem platform where my daily work includes to develop and unit test REST API's using Flask and ORM using SQLAlchemy.<br><br>- Contributed 90% of the API's for mobile application which was build in just a month keeping in mind all the secure code practices of OWASP Top 10.<br><br>- Building REST API using Flask framework, for web app & checking the load balancing of concurrent users.<br>        <br>- Building database model using SQL-Alchemy, and connecting it to Heroku.<br><br>- Security Testing of the API's !\n  Metta Social\n  metta-social\n  IT Services and IT Consulting\n  Building world’s largest common good platform to enable sustainable impact at scale!\n  4002\n\nCompany Role #5:\n  Research And Development Engineer\n  2022-01 - N/A\n  N/A\n  Cerence Inc.\n  cerence\n  Software Development\n  Cerence is the global industry leader in creating unique, moving experiences for the mobility world.\n  44039\n\n>>> SCHOOL ROLES <<<\nNo school roles listed.",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 512]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

<!-- ### Direct Usage (Transformers) <details><summary>Click to see the direct usage in Transformers</summary> </details> --> <!-- ### Downstream Usage (Sentence Transformers) You can finetune this model on your own dataset. <details><summary>Click to expand</summary> </details> --> <!-- ### Out-of-Scope Use *List how the model may foreseeably be misused and address what users ought not to do with the model.* -->

Evaluation

Metrics

Binary Classification

• Datasets: ``, rec_test_evaluator and query_test_evaluator
• Evaluated with <code>custom_evaluator.BinaryClassificationEvaluator</code>

<!-- ## Bias, Risks and Limitations *What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.* --> <!-- ### Recommendations *What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.* -->

Training Details

Training Dataset

Unnamed Dataset

• Size: 155,651 training samples
• Columns: <code>text1</code> and <code>text2</code>
• Approximate statistics based on the first 1000 samples:

  	text1	text2
  type	string	string
  details	<ul><li>min: 5 tokens</li><li>mean: 231.14 tokens</li><li>max: 2048 tokens</li></ul>	<ul><li>min: 51 tokens</li><li>mean: 618.32 tokens</li><li>max: 2048 tokens</li></ul>

• Samples:

  text1	text2
  <code>=== PERSON ===<br>Dr. Amol Charegaonkar<br>Alumnus of IIM Indore (FDP-Executive Education)<br>Specialties: Data Analysis, Finance<br>Pune, Maharashtra, India<br>4511<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #1:<br> Principal Consultant<br> 2015-04 - 2021-06<br> N/A<br> Maruma Consultancy<br> maruma-consultancy<br> Human Resources Services<br> People Development Simplified<br> 218<br><br>Company Role #2:<br> Assistant Manager – Credit (Policy - Unsecured loans)<br> 2008-06 - 2009-01<br> Job Responsibilities: As a Team Member of the Credit Policy & Risk function<br>1.Analysis of Delinquent cases in Personal loans and Business Loans<br>2.Online analysis of cases booked under Personal and Business Loans <br>3.Auditing Pan India cases for credit underwriting w. r. t. Policy <br>4.Analyzing and publishing a Monthly Portfolio Review for Personal Loans from the Risk Perspective like Delinquency & Bounce Trend and Credit Appraisal Process <br>5.Incorporate the necessary changes required in the Lending Policy consid...</code>	<code>=== PERSON ===<br>Rajesh Baj<br>N/A<br>Hyderabad, Telangana, India<br>491<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #1:<br> Sr.Officer Finance and Accounts<br> 2010-10 - N/A<br> N/A<br> JK Agri Genetics Limited<br> jk-agri-genetics-limited<br> Biotechnology Research<br> Harvesting Happiness Through Innovations <br> 107873<br><br>Company Role #2:<br> Assistant Manager<br> 2013-10 - N/A<br> N/A<br> JK Agri Genetics Limited<br> jk-agri-genetics-limited<br> Biotechnology Research<br> Harvesting Happiness Through Innovations <br> 107873<br><br>Company Role #3:<br> Officer- Finance and Accounts<br> 2007-05 - 2007-07<br> Market Research in following Areas:-<br>1. Bandra-Kurla complex<br>2. Solitaire corporate Park Andheri<br>3. Udyog Bhavan, Garegaon<br> HSBC<br> hsbc<br> Financial Services<br> N/A<br> N/A<br><br>>>> SCHOOL ROLES <<<<br>No school roles listed.</code>
  <code>=== PERSON ===<br>Gabriela Menezes<br>• Experiência em localização Brasil - TAXBRA, Nota Fiscal Eletrônica.<br>• Conhecimento em SAP Activate e ASAP.<br>• Experiência em suporte ao Cliente - AMS<br>São Paulo, São Paulo, Brazil<br>454<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #1:<br> Consultora SAP SD S/4HANA<br> 2023 - N/A<br> Atuação em - AMS <br>Analise e resolução de chamados dentro do SLA estabelecido.<br>Atuação nas diversas solicitações de melhoria.<br>Resolução de problemas relacionados a NF-e<br>Testes unitários e integrados<br>Configuração de impostos e leis fiscais.<br>Configuração de conta contábil.<br>Resolução de erros relacionados a Pricing.<br>Conhecimento e vivência no processo de Retorno Simbólico.<br>Apoio as frentes MM e FI<br>Apoio a equipe de projetos em configurações, testes unitários e integrados.<br> ITGCON Integração e Consultoria em Sistemas <br> itgcon<br> IT Services and IT Consulting<br> Mais que uma implementadora, somos a ITGCON consultoria.<br> 4116<br><br>Company Role #2:<br> Co...</code>	<code>=== PERSON ===<br>Rafael Menezes<br>Consultor funcional SD com experiência em AMS, formado em Tecnologia da Informação.<br>- Principais qualificações:<br>- Conhecimento em Localização Brasil, Pricing, Nota Fiscal Eletrônica e GRC.<br>- Conhecimento das principais tabelas do módulo SD.<br>- Conhecimento na configuração de impostos na J1BTAX<br>- Conhecimento no cadastro de dados mestres de clientes e<br>materiais.<br>São Paulo, Brazil<br>117<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #1:<br> Consultor SAP SD Pleno<br> 2023-05 - N/A<br> N/A<br> Capgemini<br> capgemini<br> IT Services and IT Consulting<br> Get the future you want<br> 6721701<br><br>Company Role #2:<br> Consultor Funcional SD JR/PL<br> 2021-12 - 2023-01<br> - Suporte ao cliente – AMS;<br>- Análise e resolução de chamados respeitando SLA;<br>- Testes em cenários de vendas;<br>- Configuração de impostos – J1BTAX;<br>- Resolução de erros referentes a rejeição de NF-e<br>- Resolução de erros referentes a CFOP<br>- Resolução de erros referentes a contas contábeis<br>- Criação de...</code>
  <code>=== PERSON ===<br>Anastasia Hanan, M.S., M.B.A (she/her)<br>~10 years of user research in both startups and enterprise. I am a senior researcher passionate about leading research that grounds product vision in data and emotion, highlights the user voice along with immediate/long term project risks, and optimizes for radically collaborative tactical execution through ambiguous problem spaces. <br><br>Specialties: Complex systems research, moving between qualitative research to quantitative insights at scale, crypto, Qualtrics, innovation and strategy, process optimization, rapid insight, cross functional collaboration, making AI make sense, visual storytelling, presentations that win contests, get millions in funding, and/or create actionable product roadmaps.<br>United States<br>1101<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #1:<br> UX Researcher<br> 2022-11 - N/A<br> N/A<br> United States Digital Service<br> united-states-digital-service<br> Government Administration<br> We're mission-driven professio...</code>	<code>=== PERSON ===<br>Stephanie Kokotakis<br>N/A<br>Los Angeles, California, United States<br>325<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br>No company roles listed.<br><br>>>> SCHOOL ROLES <<<<br>No school roles listed.</code>

• Loss: <code>MultipleNegativesSymmetricRankingLoss</code> with these parameters:

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

Evaluation Dataset

Unnamed Dataset

• Size: 47,595 evaluation samples
• Columns: <code>text1</code>, <code>text2</code>, and <code>label</code>
• Approximate statistics based on the first 1000 samples:

  	text1	text2	label
  type	string	string	int
  details	<ul><li>min: 5 tokens</li><li>mean: 221.5 tokens</li><li>max: 2048 tokens</li></ul>	<ul><li>min: 50 tokens</li><li>mean: 605.95 tokens</li><li>max: 2048 tokens</li></ul>	<ul><li>0: ~82.40%</li><li>1: ~17.60%</li></ul>

• Samples:

  text1	text2	label
  <code>Nonprofit and association management expert</code>	<code>=== PERSON ===<br>Marius Strydom<br>Marius Strydom has spent 20 years producing innovative research on the SA financial services industry, with a specific focus on insurance. He was rated the top SA life assurance analyst twice and the top short-term insurance analyst 5 times while working at Investec and BJM. At Bank of America Merrill Lynch, Marius led the EEMEA Financial Sector Research team to a second place in the 2013 Institutional Investor survey. During 2013 and 2014, he worked on academic research, culminating in being published in the South African Actuarial Journal. Since 2015, as part of his work with MLAX Consulting, Marius analysed and produced bespoke research on numerous SA and UK financial companies. In 2020, Marius founded Austin Lawrence Gidon and together with global partner, Edison Investment Research, they are driving the sponsored research revolution in South Africa.<br>City of Cape Town, Western Cape, South Africa<br>1351<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #...</code>	<code>0</code>
  <code>Expert in Alienware product support</code>	<code>=== PERSON ===<br>Ryan Talbot<br>With over 25 years of experience in physical security, Identity management, and video surveillance, Ryan is recognized as a leader in the design and implementation of major technology solutions.<br><br>Ryan has managed major security and IT system implementations for government, commercial, transport, data centres, gaming & casinos, retail, law enforcement and critical infrastructure environments. Ryan has worked closely with major technology brands in the design and development of hardware and software products throughout; Australia, Europe, North America and Asia.<br><br>A leader in advanced video solutions and a former committee member of the AS/NZS 62676:5 standards, an experienced project manager and is PRINCE2 certified. He has a certificate IV in security risk management and qualifications in security electronics and sports performance.<br><br>Ryan has also been involved in the development and design of many new technologies for IP cameras, customised ...</code>	<code>0</code>
  <code>=== PERSON ===<br>Kevin Watson<br>Specialties: Parametric cost estimation,<br>WLC / LCC analysis, <br>Logistics and supportability modelling. <br>Economic LORA, <br>Spares modelling,<br><br>Tools / Applications: PRICE® TruePlanning®, EDCAS®, VMetric®<br>United Kingdom<br>389<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #1:<br> Senior Consultant<br> 2013-09 - N/A<br> N/A<br> QinetiQ<br> qinetiq_2<br> Defense and Space Manufacturing<br> Create It. Test It. Use It.<br> 122597<br><br>Company Role #2:<br> Support Modelling Team Lead<br> 2004-06 - 2011-02<br> N/A<br> General Dynamics UK Limited<br> general-dynamics-uk-limited<br> Defense and Space Manufacturing<br> General Dynamics UK is one of the UK’s leading defence companies and an important supplier to the UK Ministry of Defence<br> 29195<br><br>Company Role #3:<br> Estimation Specialist Senior Engineer<br> 2011-02 - 2013-09<br> N/A<br> General Dynamics UK Limited<br> general-dynamics-uk-limited<br> Defense and Space Manufacturing<br> General Dynamics UK is one of the UK’s leading defence com...</code>	<code>=== PERSON ===<br>Kyle Biron, M.S.<br>N/A<br>Raleigh-Durham-Chapel Hill Area<br>308<br><br>=== ROLES ===<br>>>> COMPANY ROLES <<<<br><br>Company Role #1:<br> Clinical Informatics Analyst Associate<br> 2017-05 - 2018-05<br> • Data analyst at the Grow Baby Grow biostatistics lab at Kennesaw State University <br>• Statistical programming and statistical modeling using the Pediatrix Medical Group, Inc. database of over 1.2 million infant records to solve medical problems<br>• Received acknowledgement in a Neonatology publication comparing BMI with other proportionality measurements<br>• En route to another publication comparing popular intrauterine preterm growth curves<br>• Supervisory experience by managing and collaborating with other biostatisticians<br> Kennesaw State University<br> kennesaw-state-university<br> N/A<br> We are KSU Owls. And together, we're ascending.<br> 167722<br><br>Company Role #2:<br> Health Data Analyst II<br> 2022-06 - N/A<br> N/A<br> Nuna Inc.<br> nuna-inc<br> IT Services and IT Consulting<br> Driven by Data, Guided by Com...</code>	<code>0</code>

• Loss: <code>MultipleNegativesSymmetricRankingLoss</code> with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 32
• learning_rate: 2e-05
• num_train_epochs: 1
• warmup_ratio: 0.1
• bf16: True
• load_best_model_at_end: True
• batch_sampler: no_duplicates

All Hyperparameters

<details><summary>Click to expand</summary>

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• 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: 1
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 2e-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: 0.1
• 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: 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}
• 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: no_duplicates
• multi_dataset_batch_sampler: proportional

</details>

Training Logs

<details><summary>Click to expand</summary>

• The bold row denotes the saved checkpoint.

</details>

Framework Versions

• Python: 3.12.8
• Sentence Transformers: 3.4.1
• Transformers: 4.48.1
• PyTorch: 2.6.0+cu124
• Accelerate: 1.3.0
• Datasets: 3.2.0
• 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",
}

<!-- ## Glossary *Clearly define terms in order to be accessible across audiences.* --> <!-- ## Model Card Authors *Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.* --> <!-- ## Model Card Contact *Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.* -->