This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for retrieval.

Model Details

Model Description

Model Type: Sentence Transformer
Base model: BAAI/bge-base-en-v1.5
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Supported Modality: Text

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'BertModel'})
  (1): Pooling({'embedding_dimension': 768, 'pooling_mode': 'cls', '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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("NaqibL/bge-base-sgmarket-v2")
# Run inference
sentences = [
    'Data Engineer\n\nOur client, one of Asia-Pacific’s leading organizations is looking for: Data Engineer Scope of Work: Key Responsibilities Design, develop and deploy data tables, views and marts in data warehouses, operational data store, data lake and data virtualization.\nPerform data extraction, cleaning, transformation, and flow.\nDesign, build, launch and maintain efficient and reliable large-scale batch and real-time data pipelines with data processing frameworks.\nIntegrate and collate data silos in a manner which is both scalable and compliant Collaborate with Product Manager, Data Architect, Business Analysts, Frontend Developers, Designers and Data Analyst to build scalable data-driven products.\nBe responsible for developing backend APIs & working on databases to support the applications.\nWork in an Agile Environment that practices Continuous Integration and Delivery.\nWork closely with fellow developers through pair programming and code review process.\nQualifications: Familiar with GIS platforms: ArcGIS Server, PostGIS Able to use spatial Python libraries: GeoPandas, Shapely Build scalable geospatial data pipelines for ingestion, transformation, storage, and quality checks.\nImplements data cataloging, metadata, lineage, and security; optimizes storage and performance for GIS analytics (e.g., PostGIS, data lakes).\nProvides reliable data services and contracts to support downstream GIS analytics and front-end visualizations, collaborating closely with GIS and frontend teams.\nData management: applies validation, cleansing, reprojection, metadata, and data lineage of geospatial data Designs enterprise geospatial data models, schemas, pipelines, and services; builds or configures web map services and front-end visualization components.',
    "Data Engineer (IT Consultancy, AWS)\n\nYour new company An IT solutions consultancy is looking for an experienced Data Engineer to join their team. Your new role Build and maintain scalable data systems including databases, warehouses, and processing architectures. Design data pipelines for modeling, mining, and production to support analytics and business needs. Clean and prepare raw data using programming tools for descriptive and predictive modeling applications. Optimize data quality, reliability, and efficiency through strategic improvements and system Organize data assets and catalogs for efficient storage, access, and retrieval, tune SQL performance What you'll need to succeed Over 3 years’ experience in data architecture, warehousing, modeling, ETL/ELT, and streaming solutions. Skilled in Kubernetes-based DevOps, container orchestration, CI/CD pipelines, and microservices deployment. Proficient in Oracle SQL/PLSQL development and basic understanding of Oracle database architecture. Experienced in AWS services including EC2, S3, EMR, Redshift, Athena, and Kinesis. Strong programming skills in Python, R, SQL; familiar with Java, Spark, Hive, and Airflow. Hands-on with data crawling, modeling, lake formation, warehouse construction, and machine learning deployment. What you'll get in return In return, you'll be part of an organization that values its employees. You’ll be rewarded with: Structured career growth and plenty of developmental opportunities Hybrid and stable working environment What you need to do now If you're interested in this role, click 'apply now' or for more information and a confidential discussion on this role or to find out about more opportunities in Technology, contact Yuki Cheung or email yuki.cheung@hays.com.sg. Referrals are welcome. At Hays, we value diversity and are passionate about placing people in a role where they can flourish and succeed. We actively encourage people from diverse backgrounds to apply. EA Reg Number: R22110258 | EA License Number: 07C3924 | Company Registration No: 200609504D",
    'Research Assistant\n\nJob Scope: Gather and analyse data and feedback on the research evaluation projects (like Mid-Term reviews and Full-Term reviews) through various modalities such as engaging schools and stakeholders, as well as developing / managing surveys and focused group discussions / interviews.\nCode and verify quantitative and qualitative data in accordance with specified research protocols and coding procedures.\nDocument key processes of research / evaluation work, monitor progress and provide regular updates of projects Requirements: Trained in Data Science and possess working knowledge of Statistics, Measurement and Research.\nAt least 1 year of relevant experience in conducting research and data analytics.\nProficient in Microsoft Office Suite of tools, especially MS Excel.\nFamiliarity with analytical and statistical software and database management such as Tableau.\nStrong conceptual, analytical and problem-solving skills.\nMeticulous and possess leadership qualities to lead and drive results.\nAbility to work well independently and in teams and within tight timelines.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.8860, 0.6206],
#         [0.8860, 1.0000, 0.6310],
#         [0.6206, 0.6310, 1.0000]])

Evaluation

Metrics

Triplet

Dataset: eval
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.8534

Training Details

Training Dataset

Unnamed Dataset

Size: 26,953 training samples
Columns: sentence_0, sentence_1, and sentence_2

Approximate statistics based on the first 1000 samples:

	sentence_0	sentence_1	sentence_2
type	string	string	string
details	min: 31 tokens mean: 305.72 tokens max: 512 tokens	min: 45 tokens mean: 309.59 tokens max: 512 tokens	min: 41 tokens mean: 291.49 tokens max: 512 tokens

Samples:

sentence_0	sentence_1	sentence_2
Customer Service Officer (6 Months Contract) \| Residential[object Object][object Object]Our client is in the Property Mangement Industry.[object Object]Due to business needs, they are now recruiting for Customer Service Officer to be part of their team for ongoing transformation projects.[object Object]They are located at Bugis - Walking distance from the MRT station.[object Object]5.25 days work week - Mon-Fri: 9am-6pm, Alt Sat: 9am-1pm Duties Coordinate appointment schedules with residents.[object Object]Ensure workers report punctually for assigned duties.[object Object]Inform residents promptly in the event of any delay in scheduled appointments.[object Object]Monitor and track the progress of ongoing work assignments.[object Object]Requirements Minimum qualification: O Level or equivalent.[object Object]Excellent communication skills.[object Object]Interested candidates who wish to apply for the advertised position, please click APPLY NOW or email an updated copy of your resume/cv.	Auto Parts CS Coordinator (Tuas, 5 Days, $3.5k)[object Object][object Object]Salary: up to $3,500 Work Days: 5 Days (9am to 6.15pm) Work Location: Tuas (near MRT) Responsibilities: - Sales Order Processing (prepare quotation, sales order, invoice, D/O & P/O) for domestics and export markets.[object Object]- Coordinate packing and delivery arrangements for orders.[object Object]- Follow up on customer enquiries and quotations.[object Object]- Liaise with internal departments to ensure smooth operations and customer support.[object Object]- Monitor inventory movement and maintain records.[object Object]- Assist with general administrative duties as assigned.	`10 weeks Temp PSA Patient Service Associate (For Call Centre) [object Object]`

Commitment period: minimally till early September Location : Jalan Bukit Merah Central Working Hours: Office hours 5.5days Mon to Fri - 8am to 5pm Every Sat - 8am to 12pm Salary: $10.50 / hour Job Scope - Answer and manage appointments calls and general enquiries from members of public and patients with high service quality and operational standards.

Assist to answer calls and provide one-stop information and assistance to callers when required.
Handle the feedbacks from patients and resolving service related incidents, as well as enquiries such as the rates of consultations and medical package.
Liaise with other departments on the Call Centre Services and procedures - Work with the Call Centre Executive to maintain a high service level of Call Centre Services, and update patients' data when required.
Perform other ad-hoc duties as assigned Requirement: - Min N / O / A / NITEC in any field - No experience req... | | [object Object] 1 Year Call Centre Officer (Up to $3,000) #NJA[object Object][object Object]Ubi Any 5 days of the week from Mon to SUN 8am to 6pm (Mon-Fri), 8am to 5pm (Sat/SUN) Attend to incoming calls Provide customer service Escalate matters to relevant personnel Other ad hoc duties assigned Interestedapplicants can send their detailed resumes to [object Object] call JANE @ 6735 1955.[object Object]JANE LUI JIE'EN CEI: R1104482 Company Reg.No.[object Object]199601303W || EA Licence No.[object Object]99C4599 | Customer Service Officer (Live Chat) (Insurance) (Up to $4,000) (12 Months Contract) #NKC[object Object][object Object]Attend to customer and financial representative enquiries via phone and live chat professionally and promptly.[object Object]Resolve complaints and issues at the first touchpoint wherever possible.[object Object]Maintain accurate records of interactions in the CRM system.[object Object]Ensure timely follow-up on enquiries and escalate when necessary.[object Object]Assess appeals and provide recommendations to management.[object Object]Support additional duties during operational needs or peak periods.[object Object]Interested applicants may email resume to [object Object] Chooi Kelly (CEI Registration No: R25136207) Recruit Express Pte Ltd (EA: 99C4599) | 10 weeks Temp PSA Patient Service Associate (For Call Centre) [object Object]


[object Object]


[object Object]
[object Object]
[object Object]

Perform other ad-hoc duties as assigned Requirement: - Min N / O / A / NITEC in any field - No experience req... | | CUSTOMER SERVICE[object Object][object Object]Ensure customer bookings are promptly documented, processed and reviewed for accuracy and completeness. Input export job reference. Any special shipment requirements shall be resolved with the shipper prior accepting the booking. Keep Sales Personnel about their bookings. Upon receipt of booking from shipper, Customer Service will book shipment direct with shipping lines or our consol for both FCL and LCL cargo. After confirmation of space with shipping lines or consol, Customer Service will advise shipper via email or fax. Customer Service will proceed to arrange the trucking and collection of cargo if customer require this service. Any changes in vessel details or delay in arrival date will made known to shipper via phone or email by Customer Service. Ensure all cargoes send in good condition and if any damage shall revert to customer immediately. Verify vendor’s invoice and close files. Other ad-hoc duties as assigned by the supervisor | Studio Ambassador (am Pilates)

As our Studio Ambassador , you have an influence over our members by fostering a positive atmosphere, building community engagement, promoting membership growth.
We are seeking driven and customer-oriented individuals who are willing to go to the extra mile for our members.
Join us and be part of a dynamic and upbeat team that will value add your learning and growth as a Studio Ambassador !
Your contributions to our joint success: Welcome members to the studio and provide exceptional customer service Manage studio operations, including scheduling, check-ins, and maintaining a clean and organized environment Promote studio services, driving membership sales and retention Engage with the local community through social gatherings, workshops, and member appreciation events to strengthen relationships Contribute ideas and collect feedback to improve the overall client experience and studio atmosphere What does it take to be a Studio Ambassador: Passion for he...

| Service Desk Agent ( Interns)

We are hiring Internship, you will be equipped with the necessary skillset and knowledge to better prepare you in securing your ideal job, especially in the digital space.
Responsibilities: To log, validate and diagnose customer issues, on the full range of products and applications used at the customer site.
Providing the customer with a solution through information gathering, analytical trouble shooting and problem research, or to route or escalate the call to the appropriate resolution group.
Ensure escalation and management of calls is to agreed service levels.
Requirement Proficient in Microsoft Office.
Knowledge in desktop operating systems, various software applications and basic hardware for the PC; principles and theories of network systems and Internet technology.
Good oral and written communication skills and ability to establish and maintain effective working relationship.
Ability to coordinate several diverse job requirements and projects.
Ab...

Loss: CachedMultipleNegativesRankingLoss with these parameters:

json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim", 4 "mini_batch_size": 32, 5 "gather_across_devices": false, 6 "directions": [ 7 "query_to_doc" 8 ], 9 "partition_mode": "joint", 10 "hardness_mode": null, 11 "hardness_strength": 0.0 12}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 64
per_device_eval_batch_size: 64
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

do_predict: False
eval_strategy: no
prediction_loss_only: True
per_device_train_batch_size: 64
per_device_eval_batch_size: 64
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: 3
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_ratio: None
warmup_steps: 0
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
enable_jit_checkpoint: False
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
use_cpu: False
seed: 42
data_seed: None
bf16: False
fp16: False
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: -1
ddp_backend: None
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: False
ignore_data_skip: False
fsdp: []
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
parallelism_config: None
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch_fused
optim_args: None
group_by_length: False
length_column_name: length
project: huggingface
trackio_space_id: trackio
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
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: None
hub_always_push: False
hub_revision: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_for_metrics: []
eval_do_concat_batches: True
auto_find_batch_size: False
full_determinism: False
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
include_num_input_tokens_seen: no
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: True
use_cache: False
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	eval_cosine_accuracy
1.0	211	-	0.7835
2.0	422	-	0.8419
2.3697	500	3.0351	-
3.0	633	-	0.8534
-1	-1	-	0.8534

Training Time

Training: 4.7 hours

Framework Versions

Python: 3.12.13
Sentence Transformers: 5.4.0
Transformers: 5.0.0
PyTorch: 2.10.0+cu128
Accelerate: 1.13.0
Datasets: 4.8.5
Tokenizers: 0.22.2

Citation

BibTeX

Sentence Transformers

bibtex

@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

bibtex

@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}
}