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 semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

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

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': 512, 'do_lower_case': True}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("qnguyenle/bge-it-helpdesk-retriever")
# Run inference
sentences = [
    "Represent this sentence for searching relevant passages: Users can't log on to Windows if a multi-app kiosk profile is assigned",
    "Table of contents\nExit editor mode\nAsk Learn\nAsk Learn\nReading mode\nTable of contents\nRead in English\nAdd\nAdd to plan\nEdit\nCopy Markdown\nPrint\nNote\nAccess to this page requires authorization. You can try\nsigning in\nor\nchanging directories\n.\nAccess to this page requires authorization. You can try\nchanging directories\n.\nUsers can't log on to Windows if a multi-app kiosk profile is assigned\nFeedback\nSummarize this article for me\nThis article helps you fix an issue in which a user can't log on to a Microsoft Entra joined Windows 10 or Windows 11 computer if a multi-app kiosk profile is assigned.\nSymptoms\nWhen a user tries to log on to a Microsoft Entra joined Windows 10 or Windows 11 computer that has a multi-app kiosk profile assigned, the attempt fails immediately before the user profile is loaded.\nIn this situation, the kiosk profile logon type is\nMicrosoft Entra user\nor\nGroup\n. Additionally, the Windows 10 or Windows 11 computer uses a local account, and you notice the following error messages in the Event Viewer logs:\nMicrosoft Entra ID - Operational logs (Sample 1 - MFA required via conditional access):\nLog Name:      Microsoft-Windows-AAD/Operational\nSource:        Microsoft-Windows-AAD\nDate:\n<Timestamp>\nEvent ID:      1098\nTask Category: AadTokenBrokerPlugin Operation\nLevel:         Error\nKeywords:      Error,Error\nUser:\n<User SID>\nComputer:\n<Computer Name>\nDescription:\nError: 0xCAA2000C The request requires user interaction.\nCode: interaction_required\nDescription: AADSTS50076: Due to a configuration change made by your administrator, or because you moved to a new location, you must use multi-factor authentication to access '00000003-0000-0000-c000-000000000000'.\nMicrosoft Entra ID - Operational logs (Sample 2 - Terms of Use (TOU) required via conditional access):\nLog Name:      Microsoft-Windows-AAD/Operational\nSource:        Microsoft-Windows-AAD\nDate:\n<Timestamp>\nEvent ID:      1098\nTask Category: AadTokenBrokerPlugin Operation\nLevel:         Error\nKeywords: ",
    "Table of contents\nExit editor mode\nAsk Learn\nAsk Learn\nReading mode\nTable of contents\nRead in English\nAdd\nAdd to plan\nEdit\nCopy Markdown\nPrint\nNote\nAccess to this page requires authorization. You can try\nsigning in\nor\nchanging directories\n.\nAccess to this page requires authorization. You can try\nchanging directories\n.\nTroubleshooting PKCS certificate deployment in Intune\nFeedback\nSummarize this article for me\nThis article gives troubleshooting guidance for several common issues when deploying Public Key Cryptography Standards (PKCS) certificates in Microsoft Intune. Before troubleshooting, ensure you've completed the following tasks, as explained in\nConfigure and use PKCS certificates with Intune\n:\nReview the\nrequirements for using PKCS certificate profiles\n.\nExport the root certificate from the Enterprise Certification Authority (CA).\nConfigure certificate templates on the certification authority.\nInstall and configure the Intune Certificate Connector.\nCreate and deploy a trusted certificate profile to deploy the root certificate.\nCreate and deploy a PKCS certificate profile.\nThe most common source of problems for PKCS certificate profiles has been with the configuration of the PKCS certificate profile. Review the profiles configuration and look for typos in server names or fully qualified domain names (FQDNs), and confirm the\nCertificate Authority\nand\nCertificate Authority Name\nare correct.\nCertification Authority\n: The internal FQDN of the Certificate Authority computer. For example, server1.domain.local.\nCertification Authority Name\n: The Certificate Authority Name as displayed in the certification authority MMC. Look under\nCertification Authority (Local)\nYou can use the\ncertutil command-line program\non the CA to confirm the correct name for the Certification Authority and Certification Authority Name.\nPKCS communication overview\nThe following graphic provides a basic overview of the PKCS certificate deployment process in Intune.\nAn Admin creates a PKCS certific",
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Training Details

Training Dataset

Unnamed Dataset

Size: 2,185 training samples
Columns: anchor, positive, and negative
Approximate statistics based on the first 1000 samples:
anchor positive negative
type string string string
details
min: 11 tokens
mean: 28.84 tokens
max: 93 tokens
min: 182 tokens
mean: 400.05 tokens
max: 512 tokens
min: 182 tokens
mean: 400.67 tokens
max: 512 tokens

Samples:

anchor	positive	negative
`Represent this sentence for searching relevant passages: Assume that you're a Microsoft cloud services admin who has Microsoft Azure Multi-Factor Authentication enabled. If you lose your phone or your phone number has changed, you can't sign in to your clou`	Table of contents[object Object]Exit editor mode[object Object]Ask Learn[object Object]Ask Learn[object Object]Reading mode[object Object]Table of contents[object Object]Read in English[object Object]Add[object Object]Add to plan[object Object]Edit[object Object]Copy Markdown[object Object]Print[object Object]Note[object Object]Access to this page requires authorization. You can try[object Object]signing in[object Object]or[object Object]changing directories[object Object].[object Object]Access to this page requires authorization. You can try[object Object]changing directories[object Object].[object Object]Can't use Azure Multi-Factor Authentication to sign in to cloud services after you lose your phone or the phone number changes[object Object]Feedback[object Object]Summarize this article for me[object Object]Original product version:[object Object]Cloud Services (Web roles/Worker roles), Microsoft Entra ID, Microsoft Intune, Azure Backup, Office 365 Identity Management[object Object]Original KB number:[object Object]2834954[object Object]Symptoms[object Object]Assume that you're a Microsoft cloud services admin who has Microsoft Azure Multi-Factor Authentication enabled. If you lose your phone or your phone number has changed, you can't sign in to your cloud services account (such as Office 365, Azure, or Microsoft Intune) because you didn't receive the text message or voice call f...	Table of contents[object Object]Exit editor mode[object Object]Ask Learn[object Object]Ask Learn[object Object]Reading mode[object Object]Table of contents[object Object]Read in English[object Object]Add[object Object]Add to plan[object Object]Copy Markdown[object Object]Print[object Object]Note[object Object]Access to this page requires authorization. You can try[object Object]signing in[object Object]or[object Object]changing directories[object Object].[object Object]Access to this page requires authorization. You can try[object Object]changing directories[object Object].[object Object]Error "We couldn't add member" when adding users to Teams[object Object]Applies to: Microsoft Teams[object Object]Feedback[object Object]Summarize this article for me[object Object]Symptoms[object Object]When you add an internal or external member to Teams, you see the following error:[object Object]"We couldn't add member. We ran into an issue. Please try again later."[object Object]However, members can be added directly to Microsoft 365 groups.[object Object]Workaround[object Object]This issue occurs when the value[object Object]UsersPermissionToReadOtherUsersEnabled[object Object]is set to[object Object]False[object Object]in Microsoft Entra ID. To correct the issue, change this setting to[object Object]True[object Object].[object Object]More information[object Object]Still need help? Go to[object Object]Microsoft Community[object Object].[object Object]Feedback[object Object]Was this page helpful?[object Object]Yes[object Object]No[object Object]No[object Object]Need help with this topic?[object Object]Want to try using Ask Learn...
`Represent this sentence for searching relevant passages: How do I resolve the AADSTS7000112 error that occurs when trying to use an application with Microsoft Entra ID?`	Table of contents[object Object]Exit editor mode[object Object]Ask Learn[object Object]Ask Learn[object Object]Reading mode[object Object]Table of contents[object Object]Read in English[object Object]Add[object Object]Add to plan[object Object]Edit[object Object]Copy Markdown[object Object]Print[object Object]Note[object Object]Access to this page requires authorization. You can try[object Object]signing in[object Object]or[object Object]changing directories[object Object].[object Object]Access to this page requires authorization. You can try[object Object]changing directories[object Object].[object Object]Error AADSTS7000112 - Application is disabled[object Object]Feedback[object Object]Summarize this article for me[object Object]Summary[object Object]This article discusses how to resolve the[object Object]AADSTS7000112[object Object]error that occurs when you try to sign in to an application that can be used together with Microsoft Entra ID.[object Object]Symptoms[object Object]When you try to sign in to an Azure application that's integrated into Microsoft Entra ID, you receive the following[object Object]AADSTS7000112[object Object]error message:[object Object]Application '[object Object]	Table of contents[object Object]Exit editor mode[object Object]Ask Learn[object Object]Ask Learn[object Object]Reading mode[object Object]Table of contents[object Object]Read in English[object Object]Add[object Object]Add to plan[object Object]Edit[object Object]Copy Markdown[object Object]Print[object Object]Note[object Object]Access to this page requires authorization. You can try[object Object]signing in[object Object]or[object Object]changing directories[object Object].[object Object]Access to this page requires authorization. You can try[object Object]changing directories[object Object].[object Object]Error 806 or 808 when you import Windows Autopilot device CSV files in Intune[object Object]Feedback[object Object]Summarize this article for me[object Object]This article helps you fix an issue where you receive the[object Object]ZtdDeviceAlreadyAssigned (806)[object Object]or[object Object]ZtdDeviceAssignedToOtherTenant (808)[object Object]error message when you import Windows Autopilot device CSV files in Microsoft Intune.[object Object]Symptoms[object Object]When you try to import a CSV file in the[object Object]Microsoft Intune admin center[object Object]([object Object]Devices[object Object]>[object Object]Windows[object Object]>[object Object]Windows enrollment[object Object]>[object Object]Devices[object Object](under[object Object]Windows Autopilot Deployment Program[object Object]) >[object Object]Import[object Object]), you receive an error message similar to these:[object Object]Error details[object Object]Device is already registered to the same Tenant.[object Object]Error code:[object Object]806 -[object Object]ZtdDeviceAlreadyAssigned[object Object]CSV li...
`Represent this sentence for searching relevant passages: If you try to open a file while working offline in this scenario, the attempt fails. The error message that's displayed in this situation varies, depending on the application. Word and Excel fail with`	Table of contents[object Object]Exit editor mode[object Object]Ask Learn[object Object]Ask Learn[object Object]Reading mode[object Object]Table of contents[object Object]Read in English[object Object]Add[object Object]Add to plan[object Object]Edit[object Object]Copy Markdown[object Object]Print[object Object]Note[object Object]Access to this page requires authorization. You can try[object Object]signing in[object Object]or[object Object]changing directories[object Object].[object Object]Access to this page requires authorization. You can try[object Object]changing directories[object Object].[object Object]Can't open files offline when you use Offline Files and Windows Information Protection[object Object]Applies to:[object Object]Supported versions of Windows Client[object Object]Feedback[object Object]Summarize this article for me[object Object]This article provides workarounds for an issue that prevents you from opening files offline when you're using the Offline Files feature together with Windows Information Protection.[object Object]Applies to:[object Object]Windows 10 - all editions[object Object]Original KB number:[object Object]3187045[object Object]Symptoms[object Object]Consider the following scenario:[object Object]You have Windows 10 installed.[object Object]Special folders (for example, Documents or Favorites) are redirected to a file share.[object Object]User data in the redirected folders is cached locally through the Offline Files feature.[object Object]Windows I...	Table of contents[object Object]Exit editor mode[object Object]Ask Learn[object Object]Ask Learn[object Object]Reading mode[object Object]Table of contents[object Object]Read in English[object Object]Add[object Object]Add to plan[object Object]Edit[object Object]Copy Markdown[object Object]Print[object Object]Note[object Object]Access to this page requires authorization. You can try[object Object]signing in[object Object]or[object Object]changing directories[object Object].[object Object]Access to this page requires authorization. You can try[object Object]changing directories[object Object].[object Object]HTTP Error 500.19 - Internal Server Error when launching the App-V Management console[object Object]Applies to:[object Object]Supported versions of Windows Client[object Object]Feedback[object Object]Summarize this article for me[object Object]This article provides a solution to an issue where launching the App-V 5.0 Management console that's installed on a drive other than the C: drive fails.[object Object]Applies to:[object Object]Windows Server 2008 R2 Service Pack 1, Windows 7 Service Pack 1[object Object]Original KB number:[object Object]2800730[object Object]Symptoms[object Object]When the App-V 5.0 Management console is installed on a drive other than the C: drive, launching the console may fail and generate the following error message:[object Object]HTTP Error 500.19 - Internal Server Error[object Object]Module IIS Web Core[object Object]Notification Unkno...

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim" 4}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 16
learning_rate: 2e-05
warmup_ratio: 0.1
fp16: True

All Hyperparameters

Click to expand

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: 8
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: 3
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: None
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
bf16: False
fp16: True
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}
parallelism_config: None
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch_fused
optim_args: None
adafactor: False
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
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
hub_revision: None
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
include_tokens_per_second: False
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
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss
0.1460	20	0.6858
0.2920	40	0.4905
0.4380	60	0.3966
0.5839	80	0.3695
0.7299	100	0.3053
0.8759	120	0.3325
1.0219	140	0.2328
1.1679	160	0.2043
1.3139	180	0.243
1.4599	200	0.2028
1.6058	220	0.1816
1.7518	240	0.1634
1.8978	260	0.2181
2.0438	280	0.1966
2.1898	300	0.1193
2.3358	320	0.1711
2.4818	340	0.1179
2.6277	360	0.1347
2.7737	380	0.1562
2.9197	400	0.1236

Framework Versions

Python: 3.12.13
Sentence Transformers: 3.3.1
Transformers: 4.57.6
PyTorch: 2.10.0+cu128
Accelerate: 1.13.0
Datasets: 4.0.0
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",
}

MultipleNegativesRankingLoss

bibtex

@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}
}

	anchor	positive	negative
type	string	string	string
details	min: 11 tokens mean: 28.84 tokens max: 93 tokens	min: 182 tokens mean: 400.05 tokens max: 512 tokens	min: 182 tokens mean: 400.67 tokens max: 512 tokens