/ Модель

Характеристики

Контекст
512 токенов
Параметры
25,9 млн
Архитектура
bert · dense
Базовая модель
all-MiniLM-L6-v2
Библиотека
sentence-transformers
Опубликована
23 дек 2024

SBERT_KFOLD_User_Portfolio_to_Job_Descriptions

ЭмбеддингиТекст → embedding3 профессии

Дообученный SBERT-кодировщик превращает портфолио и описания вакансий в векторы для поиска и сравнения по смыслу.

122скачиваний за 30 дней2лайкавеса обновлены23 дек 2024

Бенчмарки

Как модель справилась с задачами пользователей ailist и что о ней публикуют открытые лидерборды

Тесты пользователейпока нет

Никто ещё не опубликовал тест этой модели. Соберите бенчмарк из своих задач — результат появится здесь.

  1. Добавьте модель в бенчмарк

    Вместе с другими моделями категории «Эмбеддинги» — сравнение идёт внутри одной категории.

  2. Опишите свои задачи

    Напишите запросы, при необходимости приложите файлы и эталонные ответы.

  3. Запустите и опубликуйте

    Модели ответят на каждую задачу, судья оценит ответы. Опубликованный результат появится на этой странице.

Публичные оценки

Из открытых лидербордов. Сравнивайте результаты одного теста при одинаковых условиях — отсутствие оценки не значит, что модель хуже.

В подключённых источниках результатов для этой модели пока нет.

Что умеет модель

Описание составлено по карточке модели и метаданным Hugging Face

Nashhz/SBERT_KFOLD_User_Portfolio_to_Job_Descriptions — дообученный разработчиком модели энкодер предложений на базе sentence-transformers/all-MiniLM-L6-v2. Модель сопоставляет предложения и абзацы плотным векторным представлениям размерности 384; архитектура включает BERT, mean pooling и нормализацию, а сходство оценивается косинусной мерой. В метаданных указан размер 25,9 млн параметров и контекст до 512 токенов; в README для конфигурации модели указана максимальная длина последовательности 256 токенов.

Модель предназначена для оценки семантического сходства между портфолио пользователя и описаниями вакансий, а также для семантического поиска, поиска перефразировок, кластеризации и классификации текстов. Обучение проводилось на 16 682 парах текстов с оценками сходства. Базовая модель англоязычная; отдельная поддержка других языков не подтверждена. В README не указана лицензия, поэтому условия использования по карточке установить нельзя.

Сценарии применения

  • сопоставление портфолио пользователя с описаниями вакансий
  • ранжирование вакансий по семантическому сходству с профилем
  • семантический поиск по резюме и вакансиям
  • поиск похожих описаний вакансий
  • кластеризация текстов вакансий и портфолио

Ограничения

Англоязычная базовая модель и ограничение последовательности в 256 токенов по README могут снижать качество на других языках и длинных текстах. Лицензия и отдельные оценки качества для сопоставления портфолио с вакансиями в карточке не приведены.

Для каких профессий подходит3

Разделы каталога профессий, в задачах которых используется эта модель

Все профессии

Карточка модели

README от автора на Hugging Face, как он его опубликовал

Открыть на Hugging Face

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2. It maps sentences & paragraphs to a 384-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: sentence-transformers/all-MiniLM-L6-v2
  • Maximum Sequence Length: 256 tokens
  • Output Dimensionality: 384 tokens
  • Similarity Function: Cosine Similarity

Model Sources

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 256, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, '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("Nashhz/SBERT_KFOLD_User_Portfolio_to_Job_Descriptions")
# Run inference
sentences = [
    'I have 15+ years experiences with web development, machine learning engineering and product development. I also have 5+ years experiences with team management for developing new product and maintaining old products.',
    "I'm starting a web development company and need a senior WordPress developer who is proficient in PHP, JavaScript, HTML, and CSS. This role will require working closely with my designer to customize websites. Key Responsibilities - Custom theme development - Communicating with the Designer - Optimising websites for performance - Ongoing website maintenance The ideal candidate should - Have expert-level experience with custom theme development - Be eager to learn and adapt - Have a solid track record with WordPress - Know the pain points of WordPress and how to solve them - Benefit Experience with SEO Collaboration - We will be using TrelloWhatsappTeams for project management and collaboration tasks. Your ability to work as part of a team and communicate effectively will be crucial for our success. A passion for web development and a desire to be part of a growing company will make this a rewarding opportunity.",
    "Job Title Freelance Graphic Designer Monthly Deliverables Minimum 30 Creative Designs Budget 10,000 Month Job Description We are seeking a Freelance Graphic Designer to create high-quality and creative visuals for our projects monthly. The ideal candidate will have experience designing a wide range of materials, including images for digital platforms, brochures, banners, PDFs, and other print-ready files. This remote freelance role is expected to deliver 30 designs per month. If you're passionate about visual design and can consistently meet deadlines with high-quality work, we'd love to hear from you! Key Responsibilities Create 30+ designs per month, including - Social media graphics - Flyers, brochures, and pamphlets - PDF print files - Flex banners and large-scale designs Design for multiple formats Digital websocial media and print brochures, banners, etc.. - Collaborate with stakeholders to ensure designs align with the brand and project goals. - Make revisions and adjustments based on feedback. - Prepare print-ready files with accurate specifications. --- Required Skills - Proficiency in Adobe Creative Suite Photoshop, Illustrator, InDesign or equivalent tools. - Strong understanding of layout, typography, and color theory, - Experience in designing for both digital and print mediums. - Knowledge of print specifications and formats CMYK, DPI, bleed, etc.. - Ability to work independently and deliver within deadlines. --- Preferred Qualifications - Prior experience as a freelance designer or working in an agency setting. - Experience with branding projects - Strong portfolio showcasing past work. --- Compensation - 10,000 per month for a minimum of 30 imagesdesigns - Additional designs or complex projects may be compensated separately based on agreement. --- How to Apply Interested candidates should submit their portfolios and CVs this platform Please include samples of - Social media posts or marketing graphics - Print designs like brochures or banners - Any other relevant design work --- Additional Information - This is a remote freelance opportunity. - Payments will be made monthly upon submission and approval of deliverables. - Long-term collaboration opportunities available based on performance.",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

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

Training Details

Training Dataset

Unnamed Dataset

  • Size: 16,682 training samples
  • Columns: sentence_0, sentence_1, and label
  • Approximate statistics based on the first 1000 samples:
    sentence_0sentence_1label
    typestringstringfloat
    detailsmin: 4 tokensmean: 160.64 tokensmax: 256 tokensmin: 8 tokensmean: 163.14 tokensmax: 256 tokensmin: 0.27mean: 0.72max: 1.0
  • Samples:
    sentence_0sentence_1label
    Amazon eBay Tiktok Shop Amazon Services Amazon Seller Central Management A to Z Store Management A to Z Inventory Management Winning Product Sourcing Product Listing with SEO Listing With Variations Listing Optimization Title, Bullet Points & Description Optimization Images Optimization Product Launching FBA Shipment Creation more Amazon eBay Tiktok Shop Amazon Services Amazon Seller Central Management A to Z Store Management A to Z Inventory Management Winning Product Sourcing Product Listing with SEO Listing With Variations Listing Optimization Title, Bullet Points & Description Optimization Images Optimization Product Launching FBA Shipment Creation Sales Generation Dropshipping Store Design A+ Content Creation Amazon PPC Campaigns Brand Registry Trademark Registration Customer Services Management eBay Services eBay Store Management A to Z A to Z eBay Dropshipping Services Winning Products Sourcing Products listing with SEO Products listing With Variations Listings Optimization Title , Bullet Point & Description Optimization Images Optimization Keywords Optimization Sales Boost Products Ranking Hot selling product with 30 to 50 profit Competitor Analysis Orders Fulfillment Customer Services Management eBay Account Defect Removal Tax Exemption Management Setting Up Promotions Listing Templates Creation Tiktok Shop Services TikTok Shop Account Setup Product Listing Listing Optimization Keyword Research Product Hunting Competitor Analysis Campaign Management Influencer Collaboration TikTok Live Shopping Order Management Promotion Management TikTok Ads for Shop Content Creation for Shop Sales Analytics & Reporting Problem Solving & Issue Resolution Ongoing Shop OptimizationI'm seeking a skilled professional to assist with a variety of tasks including selling products from Amazon UAE to eBay UK via dropshipping, product sourcing, and full virtual assistance. Key Responsibilities - Product Searching & Listing Identify profitable products, create and optimize listings, and conduct market trend analysis. - SEO Management Oversee the search engine optimization for our listed products. - Selling & Listing Management List products on Amazon, eBay, and our website, while managing sales. Ideal Candidate - Previous dropshipping experience, particularly between Amazon and eBay, is a plus. - Strong skills in SEO, product sourcing, and virtual assistance. - Excellent understanding of market trends and product profitability. - Able to create and optimize product listings for maximum visibility and sales. This is a full-time position which requires dedication and a proactive approach. Please only apply if you have the necessary skills and experience.0.7151671051979065
    We are a group of young, energetic, creative & professional website developer, graphic designer and IT-Administrator who are devoted to implement your requirement with modern technology. Website Design - Development-Modification - Wordpress - Ecommerce - DynamicCustomized site Development Graphic Design - logo design - Brochure - Flyer - Leaflet - PDF Profile - Catalog - Greetings Card - PackageLabel Design - Business Card - Image RetouchEnhancementEditingManipulation IT-Admin Virtual Assistant - Product Listing - Site Content Management - Product Image Enhance - Data Processing - PDF conversion to WordExcel - Web Research - Data Scraping Why Choose Us o Quality Support for everyday 365 days even after project completion o We understand your requirements precisely to deliver Creative designs o 100 client satisfaction guaranteedWe are looking for a skilled and dedicated full-time web developer to join our team. The ideal candidate should have extensive experience working with WordPress, Divi, and Elementor, as well as the ability to create custom WordPress themes. Key Responsibilities Develop, maintain, and optimize WordPress websites. Customize and configure Divi and Elementor page builders to meet client needs. Create custom WordPress themes from scratch, ensuring they are optimized for performance and usability. Troubleshoot and resolve any website issues as they arise. Ensure websites are responsive and work seamlessly across all devices. Collaborate with our design and content teams to bring creative ideas to life. Stay up to date with the latest web development trends and best practices. Requirements Proven experience with WordPress, including custom theme development. Proficiency in Divi and Elementor page builders. Strong understanding of HTML, CSS, JavaScript, and PHP. Experience in responsive design and cross-browser compatibility. Ability to work independently and meet deadlines. Strong problem-solving skills and attention to detail. Excellent communication skills in English. Preferred Qualifications Experience with WooCommerce or other WordPress plugins. Familiarity with SEO best practices. Knowledge of version control systems like Git. If you are passionate about web development and want to be part of a growing team, we'd love to hear from you! Please submit your portfolio and CV for consideration.0.7487468719482422
    Hi there, I'm Priyanshu Agarwal I'm a Python expert with a diverse skillset that includes web scraping, Zoho and Tally Prime accounting, automation, and Python application building. With my strong foundation in Python, I can build and automate applications that meet your business needs, saving you time and resources. As a web scraping expert, I specialize in using Python, Selenium, BeautifulSoup4, and Python Requests to extract data from websites and web applications. I have experience in projects of varying scales, from small-scale data collection to large-scale data mining for enterprise-level clients. In addition to my technical expertise in web scraping, I have a strong background in accounting software such as Zoho and Tally Prime. I have experience in managing financial data, generating reports, and automating financial processes using these tools. I understand the importance of accurate and timely financial data in business decision-making, and I strive to ensure that my clients' financial data is organized, up-to-date, and easily accessible. With my experience in automation and Python application building, I can create custom solutions toI'm in need of a data scraping expert to assist in gathering market research data from various retail websites. The ideal freelancer for this project should have a robust experience with Python and Java, as well as proficiency in Odoo and Airtable. Experience in building microservices would be a significant advantage. Key Responsibilities - Scraping data from designated retail websites for market research purposes - Organizing and managing the gathered data in Airtable - Potential development of microservices for data handling, 8n8 Skills and Experience Required - Extensive experience in data scraping, particularly from retail websites - Proficiency in Python and Java - Experience with Odoo and Airtable - Prior experience in building microservices - Understanding of market research techniques and requirements0.747043251991272
  • Loss: CosineSimilarityLoss with these parameters:
    {
        "loss_fct": "torch.nn.modules.loss.MSELoss"
    }
    

Training Hyperparameters

Non-Default Hyperparameters

  • per_device_train_batch_size: 16
  • per_device_eval_batch_size: 16
  • num_train_epochs: 4
  • multi_dataset_batch_sampler: round_robin

All Hyperparameters

  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: no
  • prediction_loss_only: True
  • per_device_train_batch_size: 16
  • per_device_eval_batch_size: 16
  • 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
  • num_train_epochs: 4
  • 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: 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: 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: False
  • hub_always_push: False
  • gradient_checkpointing: False
  • gradient_checkpointing_kwargs: None
  • include_inputs_for_metrics: False
  • 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
  • batch_sampler: batch_sampler
  • multi_dataset_batch_sampler: round_robin

Training Logs

EpochStepTraining Loss
0.47945000.001
0.958810000.0004
1.438215000.0003
1.917520000.0003
2.396925000.0003
2.876330000.0002
3.355735000.0002
3.835140000.0002
0.47945000.0003
0.958810000.0003
1.438215000.0003
1.917520000.0003
2.396925000.0002
2.876330000.0002
3.355735000.0002
3.835140000.0002
0.47945000.0002
0.958810000.0003
1.438215000.0003
1.917520000.0002
2.396925000.0002
2.876330000.0002
3.355735000.0002
3.835140000.0001
0.47945000.0002
0.958810000.0002
1.438215000.0003
1.917520000.0002
2.396925000.0002
2.876330000.0002
3.355735000.0001
3.835140000.0001
0.47945000.0002
0.958810000.0002
1.438215000.0002
1.917520000.0002
2.396925000.0002
2.876330000.0001
3.355735000.0001
3.835140000.0001

Framework Versions

  • Python: 3.12.6
  • Sentence Transformers: 3.2.0
  • Transformers: 4.45.2
  • PyTorch: 2.4.1+cpu
  • Accelerate: 1.0.1
  • Datasets: 3.0.1
  • Tokenizers: 0.20.1

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",
}

Как запустить

На своей GPU в ailist. Цена — по текущим предложениям провайдеров

GPU в ailist

Универсальный рантайм · sentence-transformers

Можно запустить

Подбираем GPU…

Класс архитектуры
BertModel
Ревизия весов
63769e1

0 моделей