metadata

language:
  - en
library_name: sentence-transformers
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - dataset_size:100K<n<1M
  - loss:CachedMultipleNegativesRankingLoss
base_model: facebook/xlm-roberta-xl
metrics:
  - cosine_accuracy
  - dot_accuracy
  - manhattan_accuracy
  - euclidean_accuracy
  - max_accuracy
widget:
  - source_sentence: A bird flying.
    sentences:
      - A butterfly flys freely.
      - the animal is running
      - The woman is asleep.
  - source_sentence: There's a dock
    sentences:
      - The boat is sinking
      - A man is ice skating.
      - A man is sleeping.
  - source_sentence: A woman sings.
    sentences:
      - The woman is singing.
      - The girl is standing.
      - A boy is sledding
  - source_sentence: The boy scowls
    sentences:
      - The boy is smiling
      - a child eats cheerios
      - A woman is sitting.
  - source_sentence: a baby smiling
    sentences:
      - The boy is smiling
      - The woman is fishing.
      - A man is sitting down.
pipeline_tag: sentence-similarity
model-index:
  - name: SentenceTransformer based on facebook/xlm-roberta-xl
    results:
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: all nli dev
          type: all-nli-dev
        metrics:
          - type: cosine_accuracy
            value: 0.931
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.063
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 0.931
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.927
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 0.931
            name: Max Accuracy
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: all nli test
          type: all-nli-test
        metrics:
          - type: cosine_accuracy
            value: 0.939
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.056
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 0.936
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.938
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 0.939
            name: Max Accuracy

SentenceTransformer based on facebook/xlm-roberta-xl

This is a sentence-transformers model finetuned from facebook/xlm-roberta-xl on the sentence-transformers/all-nli dataset. It maps sentences & paragraphs to a 2560-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: facebook/xlm-roberta-xl
Maximum Sequence Length: 514 tokens
Output Dimensionality: 2560 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': 514, 'do_lower_case': False}) with Transformer model: XLMRobertaXLModel 
  (1): Pooling({'word_embedding_dimension': 2560, '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 = [
    'a baby smiling',
    'The boy is smiling',
    'The woman is fishing.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 2560]

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

Evaluation

Metrics

Triplet

Dataset: all-nli-dev
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.931
dot_accuracy	0.063
manhattan_accuracy	0.931
euclidean_accuracy	0.927
max_accuracy	0.931

Triplet

Dataset: all-nli-test
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.939
dot_accuracy	0.056
manhattan_accuracy	0.936
euclidean_accuracy	0.938
max_accuracy	0.939

Training Details

Training Dataset

sentence-transformers/all-nli

Dataset: sentence-transformers/all-nli at d482672
Size: 100,000 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: 10.9 tokens max: 52 tokens	min: 6 tokens mean: 13.62 tokens max: 42 tokens	min: 5 tokens mean: 14.76 tokens max: 55 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: CachedMultipleNegativesRankingLoss with these parameters:

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

Evaluation Dataset

sentence-transformers/all-nli

Dataset: sentence-transformers/all-nli at d482672
Size: 1,000 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: 20.31 tokens max: 83 tokens	min: 5 tokens mean: 10.71 tokens max: 35 tokens	min: 5 tokens mean: 11.39 tokens max: 32 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: CachedMultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
num_train_epochs: 1
warmup_ratio: 0.1
bf16: 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: 16
per_device_eval_batch_size: 16
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
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: 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_eval_metrics: False
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	all-nli-dev_max_accuracy	all-nli-test_max_accuracy
0	0	-	-	0.536	-
0.016	100	3.4661	3.3640	0.54	-
0.032	200	3.4151	3.3148	0.592	-
0.048	300	3.1576	1.8296	0.82	-
0.064	400	1.2552	0.8366	0.879	-
0.08	500	0.9262	0.8402	0.888	-
0.096	600	1.0456	0.8762	0.899	-
0.112	700	0.9238	0.7967	0.894	-
0.128	800	1.0304	0.8797	0.881	-
0.144	900	1.1126	0.7939	0.903	-
0.16	1000	1.0165	0.7052	0.908	-
0.176	1100	1.06	0.6926	0.905	-
0.192	1200	0.9598	0.6843	0.894	-
0.208	1300	0.7719	0.6999	0.894	-
0.224	1400	0.7425	0.6932	0.902	-
0.24	1500	0.8434	0.5910	0.914	-
0.256	1600	0.6589	0.6143	0.903	-
0.272	1700	0.8642	0.6560	0.906	-
0.288	1800	0.838	0.6180	0.907	-
0.304	1900	0.8458	0.6097	0.91	-
0.32	2000	0.774	0.5675	0.914	-
0.336	2100	0.63	0.5917	0.909	-
0.352	2200	0.6185	0.5801	0.911	-
0.368	2300	0.8125	0.5623	0.915	-
0.384	2400	0.8478	0.5355	0.91	-
0.4	2500	0.7576	0.5490	0.91	-
0.416	2600	0.6457	0.5197	0.913	-
0.432	2700	0.918	0.5464	0.91	-
0.448	2800	0.8934	0.5375	0.91	-
0.464	2900	0.8407	0.5175	0.915	-
0.48	3000	0.7036	0.5054	0.919	-
0.496	3100	0.7929	0.5101	0.912	-
0.512	3200	0.613	0.4948	0.914	-
0.528	3300	0.5894	0.4953	0.915	-
0.544	3400	0.5885	0.4931	0.915	-
0.56	3500	0.6328	0.4868	0.921	-
0.576	3600	0.6128	0.4768	0.919	-
0.592	3700	0.5823	0.4730	0.922	-
0.608	3800	0.5633	0.4638	0.922	-
0.624	3900	0.5729	0.4587	0.926	-
0.64	4000	0.7778	0.4590	0.925	-
0.656	4100	1.0064	0.4550	0.925	-
0.672	4200	0.9683	0.4505	0.927	-
0.688	4300	0.7752	0.4470	0.926	-
0.704	4400	0.6738	0.4421	0.927	-
0.72	4500	0.6862	0.4374	0.928	-
0.736	4600	0.7699	0.4361	0.93	-
0.752	4700	0.72	0.4329	0.931	-
0.768	4800	0.7304	0.4306	0.929	-
0.784	4900	0.6679	0.4284	0.93	-
0.8	5000	0.7516	0.4285	0.93	-
0.816	5100	0.6286	0.4266	0.931	-
0.832	5200	0.7631	0.4274	0.931	-
0.848	5300	0.6368	0.4270	0.932	-
0.864	5400	0.767	0.4268	0.932	-
0.88	5500	0.6655	0.4254	0.933	-
0.896	5600	0.6599	0.4249	0.931	-
0.912	5700	0.722	0.4250	0.933	-
0.928	5800	0.682	0.4247	0.931	-
0.944	5900	0.7892	0.4245	0.931	-
0.96	6000	0.8095	0.4245	0.931	-
0.976	6100	0.6783	0.4243	0.932	-
0.992	6200	0.0051	0.4246	0.931	-
1.0	6250	-	-	-	0.939

Framework Versions

Python: 3.9.10
Sentence Transformers: 3.0.0
Transformers: 4.41.2
PyTorch: 2.3.0+cu121
Accelerate: 0.26.1
Datasets: 2.16.1
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",
}

CachedMultipleNegativesRankingLoss

@misc{gao2021scaling,
    title={Scaling Deep Contrastive Learning Batch Size under Memory Limited Setup}, 
    author={Luyu Gao and Yunyi Zhang and Jiawei Han and Jamie Callan},
    year={2021},
    eprint={2101.06983},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}