README.md

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:1290285
  - loss:TemperatureScaledCosineSimilarityLoss
widget:
  - source_sentence: garden area
    sentences:
      - minimalistisk design
      - grönområde i närheten
      - vaaleat puusävyt
  - source_sentence: ainutlaatuinen rakennus
    sentences:
      - sovrum med parkettgolv
      - light wood flooring
      - modern fixtures
  - source_sentence: förortsområde
    sentences:
      - lush greenery
      - well
      - hyvin valaistu
  - source_sentence: single-level
    sentences:
      - lähellä luontoa
      - trappor till entrén
      - light colored sofa
  - source_sentence: gravel parking
    sentences:
      - asfalterad uppfart
      - rauhallinen naapurusto
      - asfalterad uppfart
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - pearson_cosine
  - spearman_cosine
model-index:
  - name: SentenceTransformer
    results:
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: trait eval
          type: trait-eval
        metrics:
          - type: pearson_cosine
            value: 0.7845309961698034
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.8182212433993569
            name: Spearman Cosine

SentenceTransformer

This is a sentence-transformers model trained. 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
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 384 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': 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("sentence_transformers_model_id")
# Run inference
sentences = [
    'gravel parking',
    'asfalterad uppfart',
    'rauhallinen naapurusto',
]
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

Semantic Similarity

• Dataset: trait-eval
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.7845
spearman_cosine	0.8182

Training Details

Training Dataset

Unnamed Dataset

• Size: 1,290,285 training samples
• Columns: sentence_0, sentence_1, and label
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	label
  type	string	string	float
  details	min: 3 tokens mean: 5.36 tokens max: 12 tokens	min: 3 tokens mean: 5.47 tokens max: 11 tokens	min: 0.0 mean: 0.34 max: 0.9

• Samples:

  sentence_0	sentence_1	label
  kerrostalo	puuaita	0.1
  corner unit	large yard	0.3
  easy access to highway	oma pysäköinti	0.3

• Loss: main.TemperatureScaledCosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• num_train_epochs: 5
• 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: 64
• per_device_eval_batch_size: 64
• 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: 5
• 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: 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: round_robin

Training Logs

Click to expand

Epoch	Step	Training Loss	trait-eval_spearman_cosine
0.0248	500	0.1261	-
0.0496	1000	0.1155	0.2026
0.0744	1500	0.119	-
0.0992	2000	0.1193	0.2516
0.1240	2500	0.118	-
0.1488	3000	0.1151	0.2638
0.1736	3500	0.1152	-
0.1984	4000	0.1093	0.2700
0.2232	4500	0.1077	-
0.2480	5000	0.109	0.2942
0.2728	5500	0.1015	-
0.2976	6000	0.1059	0.3171
0.3224	6500	0.1052	-
0.3472	7000	0.1042	0.3114
0.3720	7500	0.1006	-
0.3968	8000	0.1007	0.3328
0.4216	8500	0.1013	-
0.4464	9000	0.0938	0.3407
0.4712	9500	0.0971	-
0.4960	10000	0.0976	0.3430
0.5208	10500	0.0945	-
0.5456	11000	0.0937	0.3603
0.5704	11500	0.0919	-
0.5952	12000	0.0893	0.3916
0.6200	12500	0.0904	-
0.6448	13000	0.0866	0.3931
0.6696	13500	0.087	-
0.6944	14000	0.0848	0.4178
0.7192	14500	0.087	-
0.7440	15000	0.0827	0.4218
0.7688	15500	0.0851	-
0.7936	16000	0.0807	0.4414
0.8184	16500	0.0803	-
0.8432	17000	0.0803	0.4526
0.8680	17500	0.0784	-
0.8928	18000	0.0776	0.4592
0.9176	18500	0.0761	-
0.9424	19000	0.0739	0.4856
0.9672	19500	0.0724	-
0.9920	20000	0.0738	0.4928
1.0	20161	-	0.4863
1.0168	20500	0.0707	-
1.0416	21000	0.0699	0.4950
1.0664	21500	0.0637	-
1.0912	22000	0.0677	0.5000
1.1160	22500	0.0638	-
1.1408	23000	0.0653	0.5306
1.1656	23500	0.0661	-
1.1904	24000	0.0679	0.5292
1.2152	24500	0.0618	-
1.2400	25000	0.0616	0.5253
1.2648	25500	0.0626	-
1.2896	26000	0.0624	0.5422
1.3144	26500	0.0613	-
1.3392	27000	0.0623	0.5515
1.3640	27500	0.0601	-
1.3888	28000	0.0589	0.5606
1.4136	28500	0.06	-
1.4384	29000	0.0598	0.5774
1.4632	29500	0.0553	-
1.4880	30000	0.0596	0.5812
1.5128	30500	0.0547	-
1.5376	31000	0.0542	0.5956
1.5624	31500	0.0547	-
1.5872	32000	0.0545	0.6053
1.6120	32500	0.0525	-
1.6368	33000	0.0545	0.6137
1.6616	33500	0.0532	-
1.6864	34000	0.0525	0.6213
1.7112	34500	0.0529	-
1.7360	35000	0.0515	0.6152
1.7608	35500	0.0506	-
1.7856	36000	0.0502	0.6211
1.8104	36500	0.0513	-
1.8352	37000	0.0476	0.6344
1.8600	37500	0.0491	-
1.8848	38000	0.0485	0.6438
1.9096	38500	0.0488	-
1.9344	39000	0.0471	0.6591
1.9592	39500	0.049	-
1.9840	40000	0.0486	0.6692
2.0	40322	-	0.6619
2.0088	40500	0.044	-
2.0336	41000	0.0407	0.6729
2.0584	41500	0.0422	-
2.0832	42000	0.0427	0.6775
2.1080	42500	0.0425	-
2.1328	43000	0.043	0.6772
2.1576	43500	0.0426	-
2.1824	44000	0.04	0.6877
2.2072	44500	0.041	-
2.2320	45000	0.0422	0.6885
2.2568	45500	0.04	-
2.2816	46000	0.0399	0.6903
2.3064	46500	0.0414	-
2.3312	47000	0.0397	0.7008
2.3560	47500	0.0406	-
2.3808	48000	0.0395	0.7004
2.4056	48500	0.0375	-
2.4304	49000	0.0391	0.6964
2.4552	49500	0.0378	-
2.4800	50000	0.0379	0.7092
2.5048	50500	0.0394	-
2.5296	51000	0.0399	0.7122
2.5544	51500	0.0357	-
2.5792	52000	0.0362	0.7170
2.6040	52500	0.0383	-
2.6288	53000	0.0396	0.7218
2.6536	53500	0.037	-
2.6784	54000	0.0372	0.7312
2.7032	54500	0.0372	-
2.7280	55000	0.0365	0.7303
2.7528	55500	0.0364	-
2.7776	56000	0.0356	0.7366
2.8024	56500	0.034	-
2.8272	57000	0.036	0.7415
2.8520	57500	0.035	-
2.8768	58000	0.0355	0.7402
2.9016	58500	0.0331	-
2.9264	59000	0.0345	0.7475
2.9512	59500	0.0345	-
2.9760	60000	0.0348	0.7489
3.0	60483	-	0.7500
3.0008	60500	0.0351	-
3.0256	61000	0.0294	0.7537
3.0504	61500	0.0297	-
3.0752	62000	0.0301	0.7557
3.1000	62500	0.0311	-
3.1248	63000	0.0314	0.7628
3.1496	63500	0.0288	-
3.1744	64000	0.0311	0.7713
3.1992	64500	0.03	-
3.2240	65000	0.0307	0.7728
3.2488	65500	0.0321	-
3.2736	66000	0.028	0.7726
3.2984	66500	0.031	-
3.3232	67000	0.0283	0.7717
3.3480	67500	0.0302	-
3.3728	68000	0.0302	0.7808
3.3976	68500	0.0303	-
3.4224	69000	0.0285	0.7790
3.4472	69500	0.0304	-
3.4720	70000	0.0287	0.7856
3.4969	70500	0.0301	-
3.5217	71000	0.0285	0.7886
3.5465	71500	0.0295	-
3.5713	72000	0.0296	0.7899
3.5961	72500	0.0269	-
3.6209	73000	0.0278	0.7911
3.6457	73500	0.0299	-
3.6705	74000	0.0285	0.7898
3.6953	74500	0.0286	-
3.7201	75000	0.0281	0.7891
3.7449	75500	0.0308	-
3.7697	76000	0.0288	0.7893
3.7945	76500	0.0283	-
3.8193	77000	0.0264	0.7953
3.8441	77500	0.0265	-
3.8689	78000	0.0271	0.7942
3.8937	78500	0.0263	-
3.9185	79000	0.0278	0.7932
3.9433	79500	0.0258	-
3.9681	80000	0.028	0.7996
3.9929	80500	0.0284	-
4.0	80644	-	0.8017
4.0177	81000	0.0258	0.8022
4.0425	81500	0.027	-
4.0673	82000	0.0228	0.8034
4.0921	82500	0.0259	-
4.1169	83000	0.0257	0.8057
4.1417	83500	0.0248	-
4.1665	84000	0.025	0.8060
4.1913	84500	0.024	-
4.2161	85000	0.0267	0.8084
4.2409	85500	0.0244	-
4.2657	86000	0.0261	0.8058
4.2905	86500	0.0256	-
4.3153	87000	0.0252	0.8061
4.3401	87500	0.0246	-
4.3649	88000	0.0243	0.8095
4.3897	88500	0.0243	-
4.4145	89000	0.0251	0.8113
4.4393	89500	0.0247	-
4.4641	90000	0.0239	0.8109
4.4889	90500	0.0248	-
4.5137	91000	0.0235	0.8129
4.5385	91500	0.0246	-
4.5633	92000	0.0231	0.8132
4.5881	92500	0.0254	-
4.6129	93000	0.0249	0.8140
4.6377	93500	0.0229	-
4.6625	94000	0.025	0.8143
4.6873	94500	0.0244	-
4.7121	95000	0.0227	0.8158
4.7369	95500	0.0223	-
4.7617	96000	0.0232	0.8166
4.7865	96500	0.024	-
4.8113	97000	0.0243	0.8170
4.8361	97500	0.0229	-
4.8609	98000	0.0243	0.8172
4.8857	98500	0.0223	-
4.9105	99000	0.0252	0.8176
4.9353	99500	0.0242	-
4.9601	100000	0.0221	0.8182
4.9849	100500	0.022	-
5.0	100805	-	0.8182

Framework Versions

• Python: 3.13.2
• Sentence Transformers: 3.4.1
• Transformers: 4.49.0
• PyTorch: 2.7.0.dev20250221+cu128
• 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",
}