wilsonmarciliojr/matryoshka-embed-nli

SentenceTransformer based on distilbert/distilroberta-base

This is a sentence-transformers model finetuned from distilbert/distilroberta-base on the all-nli dataset. 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: distilbert/distilroberta-base
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • all-nli
• Language: en

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, 'architecture': 'RobertaModel'})
  (1): Pooling({'word_embedding_dimension': 768, '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("wilsonmarciliojr/matryoshka-embed-nli")
# Run inference
sentences = [
    'A construction worker peeking out of a manhole while his coworker sits on the sidewalk smiling.',
    'A worker is looking out of a manhole.',
    'The workers are both inside the manhole.',
]
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, 0.7319, 0.6196],
#         [0.7319, 1.0000, 0.6125],
#         [0.6196, 0.6125, 1.0000]])

Evaluation

Metrics

Semantic Similarity

• Datasets: sts-dev-768 and sts-test-768
• Evaluated with EmbeddingSimilarityEvaluator with these parameters:

  {
      "truncate_dim": 768
  }
  

Metric	sts-dev-768	sts-test-768
pearson_cosine	0.8557	0.8187
spearman_cosine	0.8616	0.8339

Semantic Similarity

• Datasets: sts-dev-512 and sts-test-512
• Evaluated with EmbeddingSimilarityEvaluator with these parameters:

  {
      "truncate_dim": 512
  }
  

Metric	sts-dev-512	sts-test-512
pearson_cosine	0.8565	0.8174
spearman_cosine	0.8629	0.8339

Semantic Similarity

• Datasets: sts-dev-256 and sts-test-256
• Evaluated with EmbeddingSimilarityEvaluator with these parameters:

  {
      "truncate_dim": 256
  }
  

Metric	sts-dev-256	sts-test-256
pearson_cosine	0.852	0.8141
spearman_cosine	0.8603	0.8328

Semantic Similarity

• Datasets: sts-dev-64 and sts-test-64
• Evaluated with EmbeddingSimilarityEvaluator with these parameters:

  {
      "truncate_dim": 64
  }
  

Metric	sts-dev-64	sts-test-64
pearson_cosine	0.8303	0.7975
spearman_cosine	0.8472	0.8227

Semantic Similarity

• Datasets: sts-dev-2 and sts-test-2
• Evaluated with EmbeddingSimilarityEvaluator with these parameters:

  {
      "truncate_dim": 2
  }
  

Metric	sts-dev-2	sts-test-2
pearson_cosine	0.3308	0.3651
spearman_cosine	0.4464	0.4468

Training Details

Training Dataset

all-nli

• Dataset: all-nli at d482672
• Size: 557,850 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 7 tokens mean: 10.38 tokens max: 45 tokens	min: 6 tokens mean: 12.8 tokens max: 39 tokens	min: 6 tokens mean: 13.4 tokens max: 50 tokens

• Samples:

  anchor	positive	negative
  A person on a horse jumps over a broken down airplane.	A person is outdoors, on a horse.	A person is at a diner, ordering an omelette.
  Children smiling and waving at camera	There are children present	The kids are frowning
  A boy is jumping on skateboard in the middle of a red bridge.	The boy does a skateboarding trick.	The boy skates down the sidewalk.

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          64,
          2
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Evaluation Dataset

all-nli

• Dataset: all-nli at d482672
• Size: 6,584 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 6 tokens mean: 18.02 tokens max: 66 tokens	min: 5 tokens mean: 9.81 tokens max: 29 tokens	min: 5 tokens mean: 10.37 tokens max: 29 tokens

• Samples:

  anchor	positive	negative
  Two women are embracing while holding to go packages.	Two woman are holding packages.	The men are fighting outside a deli.
  Two young children in blue jerseys, one with the number 9 and one with the number 2 are standing on wooden steps in a bathroom and washing their hands in a sink.	Two kids in numbered jerseys wash their hands.	Two kids in jackets walk to school.
  A man selling donuts to a customer during a world exhibition event held in the city of Angeles	A man selling donuts to a customer.	A woman drinks her coffee in a small cafe.

• Loss: MatryoshkaLoss with these parameters:

  {
      "loss": "MultipleNegativesRankingLoss",
      "matryoshka_dims": [
          768,
          512,
          256,
          64,
          2
      ],
      "matryoshka_weights": [
          1,
          1,
          1,
          1,
          1
      ],
      "n_dims_per_step": -1
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 132
• per_device_eval_batch_size: 132
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 132
• per_device_eval_batch_size: 132
• 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: 1.0
• num_train_epochs: 1
• 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: 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}
• 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: no_duplicates
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss	sts-dev-768_spearman_cosine	sts-dev-512_spearman_cosine	sts-dev-256_spearman_cosine	sts-dev-64_spearman_cosine	sts-dev-2_spearman_cosine	sts-test-768_spearman_cosine	sts-test-512_spearman_cosine	sts-test-256_spearman_cosine	sts-test-64_spearman_cosine	sts-test-2_spearman_cosine
0.0237	100	20.6039	9.0978	0.7994	0.8086	0.8065	0.8008	0.4457	-	-	-	-	-
0.0473	200	10.3336	7.3115	0.8260	0.8290	0.8287	0.8162	0.4528	-	-	-	-	-
0.0710	300	9.2079	6.8495	0.8348	0.8382	0.8373	0.8221	0.4315	-	-	-	-	-
0.0946	400	8.7776	6.7054	0.8423	0.8436	0.8414	0.8291	0.4517	-	-	-	-	-
0.1183	500	8.553	6.5356	0.8319	0.8328	0.8298	0.8128	0.4012	-	-	-	-	-
0.1419	600	8.2609	6.3721	0.8448	0.8446	0.8418	0.8279	0.4212	-	-	-	-	-
0.1656	700	8.1083	6.2386	0.8481	0.8476	0.8445	0.8303	0.4330	-	-	-	-	-
0.1893	800	8.0059	6.1114	0.8463	0.8474	0.8440	0.8295	0.4177	-	-	-	-	-
0.2129	900	7.7804	6.1065	0.8499	0.8501	0.8475	0.8325	0.4324	-	-	-	-	-
0.2366	1000	7.6856	6.0044	0.8476	0.8481	0.8453	0.8276	0.4243	-	-	-	-	-
0.2602	1100	7.486	5.9960	0.8513	0.8522	0.8499	0.8340	0.4227	-	-	-	-	-
0.2839	1200	7.4374	5.9497	0.8540	0.8547	0.8523	0.8370	0.4529	-	-	-	-	-
0.3075	1300	7.3986	5.8909	0.8524	0.8536	0.8502	0.8332	0.4419	-	-	-	-	-
0.3312	1400	7.3142	5.8699	0.8573	0.8577	0.8558	0.8403	0.4575	-	-	-	-	-
0.3549	1500	7.2417	5.8065	0.8567	0.8575	0.8552	0.8396	0.4268	-	-	-	-	-
0.3785	1600	7.1856	5.8084	0.8551	0.8562	0.8543	0.8401	0.4423	-	-	-	-	-
0.4022	1700	7.0993	5.7610	0.8589	0.8593	0.8571	0.8438	0.4239	-	-	-	-	-
0.4258	1800	6.946	5.7958	0.8560	0.8569	0.8549	0.8407	0.4126	-	-	-	-	-
0.4495	1900	7.0295	5.7326	0.8610	0.8620	0.8593	0.8444	0.4325	-	-	-	-	-
0.4731	2000	7.0014	5.7051	0.8581	0.8591	0.8566	0.8427	0.4073	-	-	-	-	-
0.4968	2100	6.9669	5.6948	0.8570	0.8584	0.8558	0.8423	0.4482	-	-	-	-	-
0.5205	2200	6.9038	5.6660	0.8586	0.8594	0.8575	0.8446	0.4390	-	-	-	-	-
0.5441	2300	6.8185	5.6741	0.8600	0.8604	0.8581	0.8448	0.4463	-	-	-	-	-
0.5678	2400	6.7464	5.6465	0.8548	0.8560	0.8536	0.8405	0.4499	-	-	-	-	-
0.5914	2500	6.7982	5.6309	0.8571	0.8576	0.8551	0.8437	0.4337	-	-	-	-	-
0.6151	2600	6.7341	5.5807	0.8587	0.8598	0.8571	0.8445	0.4372	-	-	-	-	-
0.6388	2700	6.6385	5.6211	0.8596	0.8610	0.8578	0.8442	0.4508	-	-	-	-	-
0.6624	2800	6.6346	5.5926	0.8601	0.8617	0.8590	0.8464	0.4329	-	-	-	-	-
0.6861	2900	6.5412	5.5911	0.8604	0.8613	0.8590	0.8476	0.4491	-	-	-	-	-
0.7097	3000	6.5813	5.5587	0.8614	0.8630	0.8605	0.8477	0.4588	-	-	-	-	-
0.7334	3100	6.6037	5.5550	0.8608	0.8619	0.8597	0.8466	0.4547	-	-	-	-	-
0.7570	3200	6.5861	5.5688	0.8592	0.8602	0.8569	0.8438	0.4553	-	-	-	-	-
0.7807	3300	6.5861	5.5339	0.8617	0.8629	0.8606	0.8480	0.4580	-	-	-	-	-
0.8044	3400	6.5206	5.5185	0.8612	0.8625	0.8603	0.8476	0.4583	-	-	-	-	-
0.8280	3500	6.457	5.5259	0.8603	0.8619	0.8595	0.8463	0.4608	-	-	-	-	-
0.8517	3600	6.5059	5.5100	0.8622	0.8636	0.8611	0.8474	0.4625	-	-	-	-	-
0.8753	3700	6.463	5.5012	0.8621	0.8636	0.8612	0.8482	0.4551	-	-	-	-	-
0.8990	3800	6.3619	5.5010	0.8612	0.8625	0.8600	0.8469	0.4573	-	-	-	-	-
0.9226	3900	6.4302	5.4862	0.8627	0.8641	0.8615	0.8488	0.4542	-	-	-	-	-
0.9463	4000	6.3869	5.4753	0.8616	0.8630	0.8604	0.8473	0.4436	-	-	-	-	-
0.9700	4100	6.3654	5.4740	0.8618	0.8633	0.8606	0.8477	0.4454	-	-	-	-	-
0.9936	4200	6.1764	5.4702	0.8616	0.8629	0.8603	0.8472	0.4464	-	-	-	-	-
-1	-1	-	-	-	-	-	-	-	0.8339	0.8339	0.8328	0.8227	0.4468

Framework Versions

• Python: 3.10.12
• 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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

MultipleNegativesRankingLoss

@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}
}