Daxtra/sbert-summaries-minilm-24-batch

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: 128 tokens
• Output Dimensionality: 384 tokens
• Similarity Function: Cosine Similarity

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': 128, '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})
)

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("Daxtra/sbert-summaries-minilm-24-batch")
# Run inference
sentences = [
    "- Data Management Specialist (Junior/Intermediate), requiring a Bachelor's degree in computer science, information systems, or related field.\n- Must have a solid understanding of web- and app-based platforms.\n- Required experience with SQL, Data warehousing, and Tableau.\n- Ability to analyze, interpret, and organize large data sets.\n- Proficiency in Microsoft Suite, G-Suite, Slack, Zoom, ZenDesk, and Monday.com is essential.",
    "- Leadership roles in healthcare and pharmaceuticals with extensive experience in contract negotiation and site management.\n- Parexel Site Contract Leader: Lead global CSA strategy development, budget creation, and legal drafting, ensuring compliance with ICH-GCP and local regulations.\n- Drafts, reviews, and finalizes contracts with study sites, managing budgets and negotiations, and maintaining quality standards.\n- Expertise in financial risk assessment, budgeting, and regulatory compliance with a Bachelor's in Criminal Justice.\n- Proficient in Microsoft Excel, PowerPoint, and Salesforce.com, with a background in biotechnology, clinical trials, and patient safety.",
    "- Experienced Technical Project Manager with 12 years in IT infrastructure and project management.\n- Expertise in IT project planning, budgeting, resource allocation, and stakeholder engagement.\n- Led successful IT infrastructure upgrades, including EFT and Manta application migrations to AWS.\n- Managed cloud and on-premise infrastructure projects, ensuring compliance and minimal disruption.\n- Holds PMP certification and proficiency in Agile and Waterfall methodologies.\n- Master's in Business Administration with a focus on Marketing and User Experience.\n- Strong analytical skills, stakeholder management, and continuous improvement practices.\n- Proficient in Microsoft Office, Windows Server, and Azure, with experience in networking, security, and VMware.",
]
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]

Evaluation

Metrics

Information Retrieval

• Dataset: vac-res-matcher
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@10	0.3697
cosine_precision@10	0.0737
cosine_recall@10	0.1092
cosine_ndcg@10	0.1177
cosine_mrr@10	0.1953
cosine_map@10	0.0713

Training Details

Training Dataset

Unnamed Dataset

• Size: 149,352 training samples
• Columns: sentence_0 and sentence_1
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1
  type	string	string
  details	min: 53 tokens mean: 115.57 tokens max: 128 tokens	min: 51 tokens mean: 119.82 tokens max: 128 tokens

• Samples:

  sentence_0	sentence_1
  - Solutions Architect/Snowflake position requiring 8+ years of experience in data management and implementation of Snowflake solutions. - Responsibilities include developing data models, optimizing data pipelines, and integrating with AWS/Azure PBAS. - Must possess deep knowledge of relational and NoSQL databases, SQL, and Unix Shell/Python Scripting. - Experience with data security, database optimization, and Snowflake's Resource Monitors. - Required: Bachelor's or Master's degree in Computer Science or related field. - Preferred: Management consulting experience, expertise in AI use cases, and experience with Cloud technologies. - Skills: Strong leadership, team management, and experience with globally distributed teams.	- Senior IT Consultant with 30 years of experience, specializing in enterprise architecture and solution architecture. - Currently a Senior Consultant, with significant experience in architecture roles and delivering solutions for mission-critical projects. - Expertise in Enterprise Architecture, Enterprise Consortium management, and IT Governance. - Leads projects ranging from $100M to $900M, focusing on architecture for business, data, and applications. - Designed enterprise-wide architecture models and created strategies for product development based on product architecture. - Skilled in modern technologies like Spring Framework, Kafka, Spring Cloud, and Docker, and in containerization with Kubernetes. - Holds Masters Degrees in Computer Science and Law, with a Bachelor's in Computer Sciences.
  - Global Head of Business Development - Financial Education, with a focus on rapid adaptation and leadership in Hong Kong. - Responsible for identifying leads, building pipelines, and converting prospects into clients, and managing international teams. - Implement relationship-based sales practices, nurture industry relationships, and lead global expansion efforts. - Requires a proven track record in financial services business development and experience in sales with tech, media, or administration. - Strong networking and relationship-building skills, leadership abilities across cultures, and excellent communication and negotiation skills. - Must have the right to work in the United Kingdom.	- Management Accountant with extensive experience in financial planning and analysis, specializing in the banking sector. - Leads financial planning and analysis roles, overseeing budget processes, regulatory reporting, and financial performance analysis. - Streamlined processes to reduce payment processing time by 15% and minimize penalties. - Led quarterly and year-end audits, ensuring accurate financial audits and implementing improvement suggestions. - Proficient in financial reporting, budgeting, forecasting, and variance analysis; adept at using Microsoft Excel and QuickBooks. - Strong communication and team management skills, with expertise in process improvement and strategic initiatives.
  - Customer Service Administrator position requiring experience in customer service or administration. - Key responsibilities include managing customer feedback, coordinating deliveries, processing refunds, and general office support. - Must possess strong communication, problem-solving, and attention to detail skills. - Proficiency in MS Office and CRM software required. - Ability to multitask and prioritize workload effectively. - Must have the right to work in the United Kingdom.	- Experienced Technical Services Coordinator with a strong background in administration and team collaboration, seeking to advance in London. - Manages UK technical services, handling device delivery, fault reporting, and resource planning. - Expert in Microsoft Office, maintaining databases and customer service through communication and negotiation. - Holds NVQ Level 2 and 3, with skills in GDPR compliance and financial transactions. - Proven experience in collecting evidence for claims, negotiating settlements, and managing key performance indicators. - Strong communication, organizational, and interpersonal skills, with proficiency in using initiative and effective negotiation. - Holds GCSE qualifications and relevant experience in customer service roles.

• Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 24
• per_device_eval_batch_size: 24
• num_train_epochs: 1
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 24
• per_device_eval_batch_size: 24
• 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: 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: 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
• eval_use_gather_object: False
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin

Training Logs

Epoch	Step	Training Loss	vac-res-matcher_cosine_map@10
0.0803	500	1.5369	-
0.1000	622	-	0.0697
0.1607	1000	1.2768	-
0.1999	1244	-	0.0692
0.2410	1500	1.2	-
0.2999	1866	-	0.0673
0.3214	2000	1.1463	-
0.3998	2488	-	0.0705
0.4017	2500	1.1206	-
0.4821	3000	1.1043	-
0.4998	3110	-	0.0683
0.5624	3500	1.0768	-
0.5997	3732	-	0.0700
0.6428	4000	1.0905	-
0.6997	4354	-	0.0705
0.7231	4500	1.0804	-
0.7996	4976	-	0.0699
0.8035	5000	1.0536	-
0.8838	5500	1.0352	-
0.8996	5598	-	0.0715
0.9642	6000	1.0292	-
0.9995	6220	-	0.0713
1.0	6223	-	0.0713

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.2.1
• Transformers: 4.44.2
• PyTorch: 2.4.1+cu121
• Accelerate: 0.34.2
• Datasets: 3.0.1
• Tokenizers: 0.19.1

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",
}

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}
}