Daxtra/sbert-summaries-mxbai-24-batch

SentenceTransformer based on mixedbread-ai/mxbai-embed-large-v1

This is a sentence-transformers model finetuned from mixedbread-ai/mxbai-embed-large-v1. It maps sentences & paragraphs to a 1024-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: mixedbread-ai/mxbai-embed-large-v1
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 1024 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': 1024, '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-mxbai-24-batch")
# Run inference
sentences = [
    '- Content Writer position for crafting engaging long-form content across various topics, requiring a 2:1 Degree in English History or a similar field.\n- Create captivating articles, features, and content that is informative and resonates with the target audience, utilizing research skills to ensure accuracy.\n- Manage multiple projects with strong organizational and time-management skills, ensuring clarity and accuracy in writing.\n- Excellent proofreading and editing skills to produce flawless content, with a diplomatic approach to client relationships and communication.\n- Collaborate with clients to transform visions into engaging content and conduct interviews to uncover compelling stories.\n- Must have a keen eye for detail and a passion for long-form storytelling and content orchestration.',
    '- Award-nominated journalist with over 12 years of experience in journalism, content writing, and SEO.\n- Currently a freelance journalist, specializing in various sectors including health, finance, and technology.\n- Freelance for publications like The Independent, CNN, and SELF, with notable success in SEO and content creation.\n- Strong skills in SEO, content writing, and copywriting, with a track record of high engagement and click-through rates.\n- Holds a Masters of Arts in Dramaturgy and Writing, and a BA in Media Journalism and Communications.\n- Proficient in WordPress, Umbraco, and Shopify; experienced in social media, content strategy, and email marketing.',
    '- Instructional Designer and Senior Project Manager with expertise in developing eLearning experiences and conducting needs analysis for global clients.\n- Expertise in writing storyboards and scripts, creating engaging instructional graphics and animations, and developing scenario-based eLearning content.\n- Proficient in utilizing trends and best practices in learning technologies, liaising between stakeholders, and managing eLearning projects to completion.\n- Holds a Master of Arts in Applied Linguistics and Bachelor of Arts in History, with a Certificate in Digital Media.\n- Skilled in Adobe Captivate, Figma, Microsoft Excel, PowerPoint, and Word.\n- Experience includes roles as a Teacher, Instructional Designer, and eLearning Developer.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# 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	0.3939
cosine_precision@10	0.0799
cosine_recall@10	0.129
cosine_ndcg@10	0.1319
cosine_mrr@10	0.2092
cosine_map@10	0.0808

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: 57 tokens mean: 116.11 tokens max: 128 tokens	min: 52 tokens mean: 118.95 tokens max: 128 tokens

• Samples:

  sentence_0	sentence_1
  - Staff Accountant position seeking individuals with 2-3 years of experience in AR and AP areas, ideally with a Bachelor's degree in accounting, finance, or related field, or equivalent work experience. - Responsibilities include processing e-commerce payments, handling payables and receivables, preparing financial statements, supporting tax reports, and managing close processes. - Requires knowledge of basic accounting principles, experience with general ledger functions, and proficiency in Microsoft Office, particularly Excel. - Strong communication, problem-solving, and organizational skills are essential. - Attributes include a high level of integrity, ability to multitask, and strong time management. - Regular attendance and adherence to health, safety, and environmental policies are required.	- Experienced Account Payable Specialist with 5 years in vendor management, invoice processing, and reconciliation. - Processed 120-150 invoices daily, handling 3 V-way matching, and reconciling over 100,000 accounts. - Skilled in using Oracle, Excel VLOOKUP, and SAP Concur for reconciliation and disbursements. - Master’s in Accountancy; proficient in Microsoft Excel, SharePoint, and SAP applications. - Excellent interpersonal, analytical, and organizational skills. - Detailed work in expense invoicing, payments, and communication with vendors.
  - Controls Assistant Project Manager position requires 3+ years of experience, stable work history, and a bachelor's degree in mechanical or electrical engineering. - Candidates must be familiar with AutoCAD and Visio, and have experience with BACnet and DDC controls. - Knowledge of Siemens, Johnson Controls, and similar control systems is essential. - EIT or PE license is preferred but not required.	- Systems Integrator with 3 years of experience in Greater Seattle Area, specialized in project integration and sales. - Currently a Systems Integrator at GE Cimplicity, responsible for sales and project integration. - Former Systems Integrator at Siemens Tia Portal; proficient in Siemens Hardware. - Experience in GE Proficy and Albireo Energy roles, focusing on mechanical engineering. - Skills: Project Integrator, Bachelor's degree in Mechanical Engineering, bilingual in English and Spanish. - Strong background in customer engagement and system sales.
  - Senior HVAC Service Technician, requiring 5+ years of experience in commercial HVAC service and repair. - Key responsibilities include diagnosing, repairing, and maintaining large commercial HVAC systems, interpreting blueprints and data, and providing customer education. - Must possess an EPA Universal Certification and extensive knowledge of commercial HVAC systems. - Requires proficiency in interpreting technical data and blueprints, and a local work history. - Strong communication and customer service skills essential. - Ideal for professionals in San Diego, CA; must be passionate about HVAC craft and committed to service excellence.	- Mechanical Technician with extensive experience in startup, commissioning, and mechanical trades. - Currently employed at Countywide Mechanical Systems, Inc. in San Diego County, California. - Skilled in mechanical systems commissioning and troubleshooting. - Strong problem-solving and attention to detail in complex mechanical environments. - Proficient in system design, installation, and maintenance of mechanical systems. - Holds certifications in mechanical technology and engineering. - Experienced in project management and collaborative teamwork. - Educated in mechanical engineering or a related field.

• 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	1.2875	-
0.1000	622	-	0.0833
0.1607	1000	1.1274	-
0.1999	1244	-	0.0822
0.2410	1500	1.0646	-
0.2999	1866	-	0.0793
0.3214	2000	0.9926	-
0.3998	2488	-	0.0773
0.4017	2500	0.9651	-
0.4821	3000	0.9499	-
0.4998	3110	-	0.0798
0.5624	3500	0.9098	-
0.5997	3732	-	0.0793
0.6428	4000	0.8948	-
0.6997	4354	-	0.0831
0.7231	4500	0.8962	-
0.7996	4976	-	0.0809
0.8035	5000	0.8677	-
0.8838	5500	0.8696	-
0.8996	5598	-	0.0816
0.9642	6000	0.8718	-
0.9995	6220	-	0.0808
1.0	6223	-	0.0808

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}
}