metadata

language:
  - en
library_name: sentence-transformers
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - loss:AdaptiveLayerLoss
  - loss:MultipleNegativesRankingLoss
base_model: distilbert/distilroberta-base
metrics:
  - pearson_cosine
  - spearman_cosine
  - pearson_manhattan
  - spearman_manhattan
  - pearson_euclidean
  - spearman_euclidean
  - pearson_dot
  - spearman_dot
  - pearson_max
  - spearman_max
widget:
  - source_sentence: Certainly.
    sentences:
      - '''Of course.'''
      - The idea is a good one.
      - the woman is asleep at home
  - source_sentence: He walked.
    sentences:
      - The man was walking.
      - The people are running.
      - The women are making pizza.
  - source_sentence: Double pig.
    sentences:
      - Ah, triple pig!
      - He had no real answer.
      - Do you not know?
  - source_sentence: Very simply.
    sentences:
      - Not complicatedly.
      - People are on a beach.
      - The man kicks the umpire.
  - source_sentence: Introduction
    sentences:
      - Analytical Perspectives.
      - A man reads the paper.
      - No one wanted Singapore.
pipeline_tag: sentence-similarity
co2_eq_emissions:
  emissions: 94.69690706493431
  energy_consumed: 0.24362341090329948
  source: codecarbon
  training_type: fine-tuning
  on_cloud: false
  cpu_model: 13th Gen Intel(R) Core(TM) i7-13700K
  ram_total_size: 31.777088165283203
  hours_used: 0.849
  hardware_used: 1 x NVIDIA GeForce RTX 3090
model-index:
  - name: SentenceTransformer based on distilbert/distilroberta-base
    results:
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts dev
          type: sts-dev
        metrics:
          - type: pearson_cosine
            value: 0.845554152020916
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.8486455482928023
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.8475103134032791
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.8505660318245544
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.8494883021932786
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.8526835635349959
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.7866563719943611
            name: Pearson Dot
          - type: spearman_dot
            value: 0.7816258810453734
            name: Spearman Dot
          - type: pearson_max
            value: 0.8494883021932786
            name: Pearson Max
          - type: spearman_max
            value: 0.8526835635349959
            name: Spearman Max
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts test
          type: sts-test
        metrics:
          - type: pearson_cosine
            value: 0.8182808182081737
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.8148039503538166
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.8132463174874629
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.8088248622918064
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.8148200486691981
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.8105059611031759
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.7499699563291125
            name: Pearson Dot
          - type: spearman_dot
            value: 0.7350068244681712
            name: Spearman Dot
          - type: pearson_max
            value: 0.8182808182081737
            name: Pearson Max
          - type: spearman_max
            value: 0.8148039503538166
            name: Spearman Max

SentenceTransformer based on distilbert/distilroberta-base

This is a sentence-transformers model finetuned from distilbert/distilroberta-base on the sentence-transformers/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 tokens
Similarity Function: Cosine Similarity
Training Dataset:
- sentence-transformers/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}) with Transformer model: 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("tomaarsen/distilroberta-base-nli-adaptive-layer")
# Run inference
sentences = [
    'Introduction',
    'Analytical Perspectives.',
    'A man reads the paper.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Evaluation

Metrics

Semantic Similarity

Dataset: sts-dev
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.8456
spearman_cosine	0.8486
pearson_manhattan	0.8475
spearman_manhattan	0.8506
pearson_euclidean	0.8495
spearman_euclidean	0.8527
pearson_dot	0.7867
spearman_dot	0.7816
pearson_max	0.8495
spearman_max	0.8527

Semantic Similarity

Dataset: sts-test
Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.8183
spearman_cosine	0.8148
pearson_manhattan	0.8132
spearman_manhattan	0.8088
pearson_euclidean	0.8148
spearman_euclidean	0.8105
pearson_dot	0.75
spearman_dot	0.735
pearson_max	0.8183
spearman_max	0.8148

Training Details

Training Dataset

sentence-transformers/all-nli

Dataset: sentence-transformers/all-nli at e587f0c
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: AdaptiveLayerLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "n_layers_per_step": 1,
    "last_layer_weight": 1.0,
    "prior_layers_weight": 1.0,
    "kl_div_weight": 1.0,
    "kl_temperature": 0.3
}

Evaluation Dataset

sentence-transformers/all-nli

Dataset: sentence-transformers/all-nli at e587f0c
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: AdaptiveLayerLoss with these parameters:

{
    "loss": "MultipleNegativesRankingLoss",
    "n_layers_per_step": 1,
    "last_layer_weight": 1.0,
    "prior_layers_weight": 1.0,
    "kl_div_weight": 1.0,
    "kl_temperature": 0.3
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 128
per_device_eval_batch_size: 128
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: False
per_device_train_batch_size: 128
per_device_eval_batch_size: 128
per_gpu_train_batch_size: None
per_gpu_eval_batch_size: None
gradient_accumulation_steps: 1
eval_accumulation_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
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: None
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: False
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
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_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	sts-dev_spearman_cosine	sts-test_spearman_cosine
0.0229	100	7.0517	3.9378	0.7889	-
0.0459	200	4.4877	3.8105	0.7906	-
0.0688	300	4.0315	3.6401	0.7966	-
0.0918	400	3.822	3.3537	0.7883	-
0.1147	500	3.0608	2.5975	0.7973	-
0.1376	600	2.6304	2.3956	0.7943	-
0.1606	700	2.7723	2.0379	0.8009	-
0.1835	800	2.3556	1.9645	0.7984	-
0.2065	900	2.4998	1.9086	0.8017	-
0.2294	1000	2.1834	1.8400	0.7973	-
0.2524	1100	2.2793	1.5831	0.8102	-
0.2753	1200	2.1042	1.6485	0.8004	-
0.2982	1300	2.1365	1.7084	0.8013	-
0.3212	1400	2.0096	1.5520	0.8064	-
0.3441	1500	2.0492	1.4917	0.8084	-
0.3671	1600	1.8764	1.5447	0.8018	-
0.3900	1700	1.8611	1.5480	0.8046	-
0.4129	1800	1.972	1.5353	0.8075	-
0.4359	1900	1.8062	1.4633	0.8039	-
0.4588	2000	1.8565	1.4213	0.8027	-
0.4818	2100	1.8852	1.3860	0.8002	-
0.5047	2200	1.7939	1.5468	0.7910	-
0.5276	2300	1.7398	1.6041	0.7888	-
0.5506	2400	1.8535	1.5791	0.7949	-
0.5735	2500	1.8486	1.4871	0.7951	-
0.5965	2600	1.7379	1.5427	0.8019	-
0.6194	2700	1.7325	1.4585	0.8087	-
0.6423	2800	1.7664	1.5264	0.7965	-
0.6653	2900	1.7517	1.6344	0.7930	-
0.6882	3000	1.8329	1.4947	0.8008	-
0.7112	3100	1.7206	1.4917	0.8089	-
0.7341	3200	1.7138	1.4185	0.8065	-
0.7571	3300	1.3705	1.2040	0.8446	-
0.7800	3400	1.1289	1.1363	0.8447	-
0.8029	3500	1.0174	1.1049	0.8464	-
0.8259	3600	1.0188	1.0362	0.8466	-
0.8488	3700	0.9841	1.1391	0.8470	-
0.8718	3800	0.8466	1.0116	0.8485	-
0.8947	3900	0.9268	1.1323	0.8488	-
0.9176	4000	0.8686	1.0296	0.8495	-
0.9406	4100	0.9255	1.1737	0.8484	-
0.9635	4200	0.7991	1.0609	0.8486	-
0.9865	4300	0.8431	0.9976	0.8486	-
1.0	4359	-	-	-	0.8148

Environmental Impact

Carbon emissions were measured using CodeCarbon.

Energy Consumed: 0.244 kWh
Carbon Emitted: 0.095 kg of CO2
Hours Used: 0.849 hours

Training Hardware

On Cloud: No
GPU Model: 1 x NVIDIA GeForce RTX 3090
CPU Model: 13th Gen Intel(R) Core(TM) i7-13700K
RAM Size: 31.78 GB

Framework Versions

Python: 3.11.6
Sentence Transformers: 3.0.0.dev0
Transformers: 4.41.0.dev0
PyTorch: 2.3.0+cu121
Accelerate: 0.26.1
Datasets: 2.18.0
Tokenizers: 0.19.1

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",
}

AdaptiveLayerLoss

@misc{li20242d,
    title={2D Matryoshka Sentence Embeddings}, 
    author={Xianming Li and Zongxi Li and Jing Li and Haoran Xie and Qing Li},
    year={2024},
    eprint={2402.14776},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

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}
}