InstalilyAI/sentence-transformer-synonyms

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2. 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
• Base model: sentence-transformers/all-MiniLM-L6-v2
• Maximum Sequence Length: 256 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': 256, 'do_lower_case': False, 'architecture': '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})
  (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("sentence_transformers_model_id")
# Run inference
sentences = [
    'install new Actuator Switch',
    'replace fan switch',
    'replace hi limit switch',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 384]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.8039, 0.8263],
#         [0.8039, 1.0000, 0.7918],
#         [0.8263, 0.7918, 1.0000]])

Training Details

Training Dataset

Unnamed Dataset

• Size: 78,106 training samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	label
  type	string	string	int
  details	min: 4 tokens mean: 6.97 tokens max: 16 tokens	min: 4 tokens mean: 7.08 tokens max: 19 tokens	1: 100.00%

• Samples:

  sentence1	sentence2	label
  Solid state relay contacts not making contact	Solid state relay not closing	1
  Single pole relay is stuck closed	Relay is stuck closed	1
  Relay potentially damaged due to wiring issue	Relay faulty	1

• Loss: OnlineContrastiveLoss

Evaluation Dataset

Unnamed Dataset

• Size: 818 evaluation samples
• Columns: sentence1, sentence2, and label
• Approximate statistics based on the first 818 samples:

  	sentence1	sentence2	label
  type	string	string	int
  details	min: 3 tokens mean: 5.53 tokens max: 10 tokens	min: 3 tokens mean: 5.93 tokens max: 14 tokens	0: ~50.00% 1: ~50.00%

• Samples:

  sentence1	sentence2	label
  Power Relay needs replacement	relay needs replacement	1
  Target needs replacement	ceramic tile needs replacement	1
  install new Infinite Switch	install new inf switch	1

• Loss: OnlineContrastiveLoss

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: epoch
• per_device_train_batch_size: 256
• per_device_eval_batch_size: 256
• learning_rate: 1e-05
• num_train_epochs: 4
• warmup_ratio: 0.02
• bf16: True

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: epoch
• prediction_loss_only: True
• per_device_train_batch_size: 256
• per_device_eval_batch_size: 256
• 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: 1e-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: 4
• max_steps: -1
• lr_scheduler_type: linear
• lr_scheduler_kwargs: {}
• warmup_ratio: 0.02
• 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: True
• 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
• 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: False
• 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: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss
1.0	7	-	28.7758
2.0	14	-	20.5461
3.0	21	-	14.5468
4.0	28	-	12.2789
5.0	35	-	7.4563
6.0	42	-	4.7709
7.0	49	-	3.7263
8.0	56	-	3.2799
9.0	63	-	3.4937
10.0	70	-	3.3956
11.0	77	-	3.2518
12.0	84	-	2.4912
13.0	91	-	1.7859
14.0	98	-	1.4185
14.2857	100	9.9923	-
15.0	105	-	1.4582
16.0	112	-	1.4355
17.0	119	-	1.2700
18.0	126	-	0.9766
19.0	133	-	0.9087
20.0	140	-	0.8227
21.0	147	-	0.7897
22.0	154	-	0.6956
23.0	161	-	0.7913
24.0	168	-	0.7902
25.0	175	-	0.7534
26.0	182	-	0.6562
27.0	189	-	0.5444
28.0	196	-	0.4464
28.5714	200	0.2576	-
29.0	203	-	0.4410
30.0	210	-	0.4314
31.0	217	-	0.3471
32.0	224	-	0.3472
33.0	231	-	0.3445
34.0	238	-	0.3404
35.0	245	-	0.3378
36.0	252	-	0.3370
37.0	259	-	0.3355
38.0	266	-	0.3339
39.0	273	-	0.3326
40.0	280	-	0.3328
41.0	287	-	0.3308
42.0	294	-	0.3308
42.8571	300	0.1918	-
43.0	301	-	0.3306
44.0	308	-	0.3304
45.0	315	-	0.3294
46.0	322	-	0.3295
47.0	329	-	0.3295
48.0	336	-	0.3297
49.0	343	-	0.3295
50.0	350	-	0.3295
1.0	4	-	0.1957
2.0	8	-	0.1227
3.0	12	-	0.1132
4.0	16	-	0.0720
5.0	20	-	0.0629
6.0	24	-	0.0591
7.0	28	-	0.0
8.0	32	-	0.0
9.0	36	-	0.0
10.0	40	-	0.0
11.0	44	-	0.0
12.0	48	-	0.0
12.5	50	0.441	-
13.0	52	-	0.0
14.0	56	-	0.0
15.0	60	-	0.0
16.0	64	-	0.0
17.0	68	-	0.0
18.0	72	-	0.0
19.0	76	-	0.0
20.0	80	-	0.0
21.0	84	-	0.0
22.0	88	-	0.0
23.0	92	-	0.0
24.0	96	-	0.0
25.0	100	0.3348	0.0
26.0	104	-	0.0
27.0	108	-	0.0
28.0	112	-	0.0
29.0	116	-	0.0
30.0	120	-	0.0
0.6536	200	7.707	-
1.0	306	-	0.7959
1.3072	400	6.7597	-
1.9608	600	6.4126	-
2.0	612	-	1.0730
2.6144	800	6.2623	-
3.0	918	-	1.1429
3.2680	1000	6.1651	-
3.9216	1200	6.12	-
4.0	1224	-	1.2153

• The bold row denotes the saved checkpoint.

Framework Versions

• Python: 3.11.0rc1
• Sentence Transformers: 5.0.0
• Transformers: 4.53.1
• PyTorch: 2.7.1+cu126
• Accelerate: 1.8.1
• Datasets: 3.6.0
• Tokenizers: 0.21.2

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",
}