Lauther/measuring-embeddings-v4.3

SentenceTransformer

This is a sentence-transformers model trained on the measuring-embeddings-v4 dataset. 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
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 1024 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • measuring-embeddings-v4

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': False}) with Transformer model: XLMRobertaModel 
  (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})
  (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("Lauther/measuring-embeddings-v4.3")
# Run inference
sentences = [
    'uncertainty points',
    'What is a Fluid?\nA Fluid is the substance measured within a measurement system. It can be a gas or liquid, such as hydrocarbons, water, or other industrial fluids. Proper classification of fluids is essential for ensuring measurement accuracy, regulatory compliance, and operational efficiency. By identifying fluids correctly, the system applies the appropriate measurement techniques, processing methods, and reporting standards.',
    'What is a Calibration Point?\nA Calibration Point represents a specific data entry in a calibration process, comparing an expected reference value to an actual measured value. These points are fundamental in ensuring measurement accuracy and identifying deviations.\n\nKey Aspects of Calibration Points:\n- Calibration Report Association: Each calibration point belongs to a specific calibration report, linking it to a broader calibration procedure.\n- Reference Values: Theoretical or expected values used as a benchmark for measurement validation.\n- Measured Values: The actual recorded values during calibration, reflecting the instrument’s response.\n- Errors: The difference between reference and measured values, indicating possible measurement inaccuracies.\nCalibration points are essential for evaluating instrument performance, ensuring compliance with standards, and maintaining measurement reliability.',
]
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]

Training Details

Training Dataset

measuring-embeddings-v4

• Dataset: measuring-embeddings-v4 at 1e3ca2c
• Size: 3,075 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 3 tokens mean: 7.55 tokens max: 17 tokens	min: 80 tokens mean: 180.22 tokens max: 406 tokens	min: 0.07 mean: 0.21 max: 0.95

• Samples:

  sentence1	sentence2	score
  last calibrated span	What are historical report values? These represent the recorded data points within flow computer reports. Unlike the report index, which serves as a reference to locate reports, these values contain the actual measurements and calculated data stored in the historical records. Flow computer reports store two types of data values: - Hourly data values: Contain measured or calculated values (e.g., operational minutes, alarms set, etc.) recorded on an hourly basis. - Daily data values: Contain measured or calculated values (e.g., operational minutes, alarms set, etc.) recorded on a daily basis. Each value is directly linked to its respective report index, ensuring traceability to the original flow computer record. These values maintain their raw integrity, providing a reliable source for analysis and validation.	0.1
  flow computer configuration	What is a Measurement Type? Measurement types define the classification of measurements used within a system based on their purpose and regulatory requirements. These types include fiscal, appropriation, operational, and custody measurements. - Fiscal measurements are used for tax and regulatory reporting, ensuring accurate financial transactions based on measured quantities. - Appropriation measurements track resource allocation and ownership distribution among stakeholders. - Operational measurements support real-time monitoring and process optimization within industrial operations. - Custody measurements are essential for legal and contractual transactions, ensuring precise handover of fluids between parties. These classifications play a crucial role in compliance, financial accuracy, and operational efficiency across industries such as oil and gas, water management, and energy distribution.	0.1
  uncertainty certificate number	What is an Uncertainty Composition? An Uncertainty Composition represents a specific factor that contributes to the overall uncertainty of a measurement system. These components are essential for evaluating the accuracy and reliability of measurements by identifying and quantifying the sources of uncertainty. Key Aspects of an Uncertainty Component: - Component Name: Defines the uncertainty factor (e.g., diameter, density, variance, covariance) influencing the measurement system. - Value of Composition: Quantifies the component’s contribution to the total uncertainty, helping to analyze which factors have the greatest impact. - Uncertainty File ID: Links the component to a specific uncertainty dataset for traceability and validation. Understanding these components is critical for uncertainty analysis, ensuring compliance with industry standards and improving measurement precision.	0.1

• Loss: CoSENTLoss with these parameters:

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

Evaluation Dataset

measuring-embeddings-v4

• Dataset: measuring-embeddings-v4 at 1e3ca2c
• Size: 659 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 659 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 3 tokens mean: 7.63 tokens max: 17 tokens	min: 80 tokens mean: 186.36 tokens max: 406 tokens	min: 0.07 mean: 0.2 max: 0.9

• Samples:

  sentence1	sentence2	score
  measurement system details	What is an Uncertainty Composition? An Uncertainty Composition represents a specific factor that contributes to the overall uncertainty of a measurement system. These components are essential for evaluating the accuracy and reliability of measurements by identifying and quantifying the sources of uncertainty. Key Aspects of an Uncertainty Component: - Component Name: Defines the uncertainty factor (e.g., diameter, density, variance, covariance) influencing the measurement system. - Value of Composition: Quantifies the component’s contribution to the total uncertainty, helping to analyze which factors have the greatest impact. - Uncertainty File ID: Links the component to a specific uncertainty dataset for traceability and validation. Understanding these components is critical for uncertainty analysis, ensuring compliance with industry standards and improving measurement precision.	0.15
  measurement system tag EMED-3102-02-010	What is a report index or historic index? Indexes represent the recorded reports generated by flow computers, classified into two types: - Hourly reports Index: Store data for hourly events. - Daily reports Index: Strore data for daily events. These reports, also referred to as historical data or flow computer historical records, contain raw, first-hand measurements directly collected from the flow computer. The data has not been processed or used in any calculations, preserving its original state for analysis or validation. The index is essential for locating specific values within the report.	0.24
  static pressure	What is a Meter Stream? A Meter Stream represents a measurement system configured within a flow computer. It serves as the interface between the physical measurement system and the computational processes that record and analyze flow data. Key Aspects of a Meter Stream: - Status: Indicates whether the meter stream is active or inactive. - Measurement System Association: Links the meter stream to a specific measurement system, ensuring that the data collected corresponds to a defined physical setup. - Flow Computer Association: Identifies the flow computer responsible for managing and recording the measurement system's data. Why is a Meter Stream Important? A meter stream is a critical component in flow measurement, as it ensures that the measurement system is correctly integrated into the flow computer for accurate monitoring and reporting. Since each flow computer can handle multiple meter streams, proper configuration is essential for maintaining data integrity and traceability.	0.1

• Loss: CoSENTLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• gradient_accumulation_steps: 4
• learning_rate: 2e-05
• num_train_epochs: 10
• warmup_ratio: 0.1

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 4
• per_gpu_train_batch_size: None
• per_gpu_eval_batch_size: None
• gradient_accumulation_steps: 4
• 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: 10
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• 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
• 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: None
• hub_always_push: False
• 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
• 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
• average_tokens_across_devices: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Click to expand

Epoch	Step	Training Loss	Validation Loss
0.8322	160	3.0564	-
0.8843	170	2.2963	-
0.9363	180	1.8767	-
0.9883	190	2.8634	-
1.0416	200	2.5195	-
1.0936	210	2.4094	-
1.1456	220	1.5141	-
1.1977	230	2.1366	-
1.2497	240	1.5389	-
1.3017	250	3.8265	-
1.3537	260	1.9989	-
1.4057	270	2.6037	-
1.4577	280	3.898	-
1.5098	290	2.9363	-
1.5618	300	3.3853	0.5155
1.6138	310	2.2995	-
1.6658	320	1.3945	-
1.7178	330	3.8312	-
1.7698	340	2.626	-
1.8218	350	1.5451	-
1.8739	360	1.1062	-
1.9259	370	2.6593	-
1.9779	380	1.773	-
2.0260	390	1.3937	-
2.0780	400	2.2228	-
2.1300	410	0.7027	-
2.1821	420	1.5933	-
2.2341	430	2.295	-
2.2861	440	1.042	-
2.3381	450	2.8671	0.3661
2.3901	460	1.879	-
2.4421	470	4.0556	-
2.4941	480	2.9677	-
2.5462	490	1.4443	-
2.5982	500	3.2575	-
2.6502	510	1.6124	-
2.7022	520	1.3976	-
2.7542	530	1.3161	-
2.8062	540	2.5047	-
2.8583	550	0.9757	-
2.9103	560	2.1051	-
2.9623	570	2.4919	-
3.0104	580	1.4737	-
3.0624	590	1.3318	-
3.1144	600	1.4474	0.4409
3.1664	610	2.3727	-
3.2185	620	0.6234	-
3.2705	630	1.9529	-
3.3225	640	1.5384	-
3.3745	650	1.5913	-
3.4265	660	0.6265	-
3.4785	670	2.1122	-
3.5306	680	1.8046	-
3.5826	690	0.8298	-
3.6346	700	1.4242	-
3.6866	710	1.5808	-
3.7386	720	1.1792	-
3.7906	730	2.7767	-
3.8427	740	1.7814	-
3.8947	750	0.5374	0.3227
3.9467	760	1.493	-
3.9987	770	1.8282	-
4.0468	780	1.6991	-
4.0988	790	0.7883	-
4.1508	800	0.841	-
4.2029	810	0.923	-
4.2549	820	0.3459	-
4.3069	830	2.3643	-
4.3589	840	0.9606	-
4.4109	850	0.7961	-
4.4629	860	1.749	-
4.5150	870	0.6536	-
4.5670	880	1.668	-
4.6190	890	0.5919	-
4.6710	900	1.2476	0.3258
4.7230	910	1.422	-
4.7750	920	0.8616	-
4.8270	930	0.2323	-
4.8791	940	2.7915	-
4.9311	950	0.6705	-
4.9831	960	1.7353	-
5.0312	970	1.7646	-
5.0832	980	1.4311	-
5.1352	990	0.7089	-
5.1873	1000	1.631	-
5.2393	1010	1.8051	-
5.2913	1020	0.5302	-
5.3433	1030	0.7428	-
5.3953	1040	0.5852	-
5.4473	1050	0.737	0.3283
5.4993	1060	1.492	-
5.5514	1070	0.9142	-
5.6034	1080	1.8887	-
5.6554	1090	1.1079	-
5.7074	1100	0.6984	-
5.7594	1110	1.7174	-
5.8114	1120	0.9411	-
5.8635	1130	1.286	-
5.9155	1140	2.1944	-
5.9675	1150	1.2478	-
6.0156	1160	0.7935	-
6.0676	1170	1.4886	-
6.1196	1180	1.3375	-
6.1717	1190	2.9167	-
6.2237	1200	0.3903	0.2734
6.2757	1210	1.326	-
6.3277	1220	0.3135	-
6.3797	1230	1.0881	-
6.4317	1240	1.5096	-
6.4837	1250	0.5525	-
6.5358	1260	0.3606	-
6.5878	1270	0.9334	-
6.6398	1280	0.5658	-
6.6918	1290	1.5978	-
6.7438	1300	0.4212	-
6.7958	1310	1.7793	-
6.8479	1320	1.5593	-
6.8999	1330	1.6738	-
6.9519	1340	0.3041	-
7.0	1350	0.5286	0.2737
7.0520	1360	1.7618	-
7.1040	1370	0.4629	-
7.1560	1380	0.4087	-
7.2081	1390	0.3099	-
7.2601	1400	0.6679	-
7.3121	1410	0.7688	-
7.3641	1420	1.223	-
7.4161	1430	0.8108	-
7.4681	1440	0.24	-
7.5202	1450	0.6616	-
7.5722	1460	1.5255	-
7.6242	1470	1.3865	-
7.6762	1480	0.2771	-
7.7282	1490	0.7809	-
7.7802	1500	0.2114	0.2259
7.8322	1510	1.6341	-
7.8843	1520	0.7665	-
7.9363	1530	0.7204	-
7.9883	1540	0.6557	-
8.0364	1550	2.0155	-
8.0884	1560	0.4718	-
8.1404	1570	0.1254	-
8.1925	1580	0.8067	-
8.2445	1590	0.3196	-
8.2965	1600	0.7162	-
8.3485	1610	0.1727	-
8.4005	1620	0.7634	-
8.4525	1630	0.2472	-
8.5046	1640	0.264	-
8.5566	1650	0.5994	0.1935
8.6086	1660	0.4445	-
8.6606	1670	0.9039	-
8.7126	1680	0.7927	-
8.7646	1690	0.4908	-
8.8166	1700	0.7486	-
8.8687	1710	1.377	-
8.9207	1720	1.025	-
8.9727	1730	1.1134	-
9.0208	1740	0.271	-
9.0728	1750	1.0931	-
9.1248	1760	0.7956	-
9.1769	1770	1.2794	-
9.2289	1780	0.3901	-
9.2809	1790	0.9033	-
9.3329	1800	0.4934	0.1680
9.3849	1810	0.5104	-
9.4369	1820	0.2879	-
9.4889	1830	0.6565	-
9.5410	1840	0.4523	-
9.5930	1850	0.7147	-
9.6450	1860	0.354	-
9.6970	1870	0.277	-
9.7490	1880	0.2066	-
9.8010	1890	0.6588	-
9.8531	1900	0.3789	-
9.9051	1910	0.8525	-
9.9571	1920	0.366	-

Framework Versions

• Python: 3.11.0
• Sentence Transformers: 3.4.1
• Transformers: 4.49.0
• PyTorch: 2.6.0+cu124
• Accelerate: 1.4.0
• Datasets: 3.3.2
• Tokenizers: 0.21.0

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",
}

CoSENTLoss

@online{kexuefm-8847,
    title={CoSENT: A more efficient sentence vector scheme than Sentence-BERT},
    author={Su Jianlin},
    year={2022},
    month={Jan},
    url={https://kexue.fm/archives/8847},
}