This is a sentence-transformers model finetuned from Unbabel/xlm-roberta-comet-small on the sentence-transformers/all-nli dataset. 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: Unbabel/xlm-roberta-comet-small
Maximum Sequence Length: 512 tokens
Output Dimensionality: 384 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: XLMRobertaModel 
  (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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("mics-nlp/xlm-roberta-small-all-nli-triplet")
# Run inference
sentences = [
    'a baby smiling',
    'A baby is unhappy.',
    'The dog has big ears.',
]
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

Triplet

Dataset: all-nli-dev
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.849
dot_accuracy	0.163
manhattan_accuracy	0.837
euclidean_accuracy	0.841
max_accuracy	0.849

Triplet

Dataset: all-nli-test
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.839
dot_accuracy	0.15
manhattan_accuracy	0.827
euclidean_accuracy	0.827
max_accuracy	0.839

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:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim" 4}

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:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim" 4}

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	-	-	0.541	-
0.016	100	3.5308	3.1817	0.558	-
0.032	200	3.2784	3.0406	0.597	-
0.048	300	3.113	2.7572	0.635	-
0.064	400	2.8296	2.4646	0.68	-
0.08	500	2.631	2.3583	0.676	-
0.096	600	2.3247	2.1394	0.706	-
0.112	700	2.2211	2.0201	0.711	-
0.128	800	2.1263	1.9560	0.757	-
0.144	900	2.2105	1.9074	0.748	-
0.16	1000	2.0637	1.9289	0.728	-
0.176	1100	2.1772	1.8796	0.741	-
0.192	1200	2.1518	1.8346	0.761	-
0.208	1300	1.728	1.8213	0.765	-
0.224	1400	1.8101	1.6321	0.772	-
0.24	1500	1.7516	1.5669	0.793	-
0.256	1600	1.4988	1.5538	0.8	-
0.272	1700	1.6695	1.5462	0.803	-
0.288	1800	1.5971	1.5499	0.783	-
0.304	1900	1.5614	1.5047	0.788	-
0.32	2000	1.522	1.4957	0.794	-
0.336	2100	1.3624	1.4153	0.814	-
0.352	2200	1.4773	1.4169	0.809	-
0.368	2300	1.6066	1.3697	0.813	-
0.384	2400	1.5106	1.3203	0.819	-
0.4	2500	1.4783	1.3417	0.817	-
0.416	2600	1.3696	1.2650	0.824	-
0.432	2700	1.5115	1.2779	0.829	-
0.448	2800	1.4834	1.2668	0.834	-
0.464	2900	1.4823	1.2621	0.836	-
0.48	3000	1.4163	1.2465	0.837	-
0.496	3100	1.4232	1.2475	0.837	-
0.512	3200	1.2193	1.1975	0.838	-
0.528	3300	1.2569	1.1816	0.838	-
0.544	3400	1.2988	1.1936	0.839	-
0.56	3500	1.5068	1.2213	0.835	-
0.576	3600	1.3022	1.1799	0.842	-
0.592	3700	1.3823	1.1910	0.831	-
0.608	3800	1.4224	1.1786	0.834	-
0.624	3900	1.3765	1.1541	0.843	-
0.64	4000	1.4987	1.1365	0.844	-
0.656	4100	1.7525	1.1394	0.843	-
0.672	4200	1.6013	1.1178	0.841	-
0.688	4300	1.3326	1.0959	0.846	-
0.704	4400	1.355	1.0757	0.848	-
0.72	4500	1.2834	1.0681	0.846	-
0.736	4600	1.2939	1.0696	0.85	-
0.752	4700	1.4069	1.0645	0.848	-
0.768	4800	1.4503	1.0609	0.849	-
0.784	4900	1.2833	1.0587	0.847	-
0.8	5000	1.3321	1.0563	0.849	-
0.816	5100	1.3006	1.0539	0.847	-
0.832	5200	1.4332	1.0527	0.847	-
0.848	5300	1.3101	1.0505	0.848	-
0.864	5400	1.3658	1.0523	0.849	-
0.88	5500	1.353	1.0520	0.849	-
0.896	5600	1.2429	1.0521	0.848	-
0.912	5700	1.3512	1.0505	0.848	-
0.928	5800	1.2995	1.0501	0.848	-
0.944	5900	1.3514	1.0491	0.849	-
0.96	6000	1.3976	1.0490	0.848	-
0.976	6100	1.2112	1.0487	0.848	-
0.992	6200	0.0033	1.0492	0.849	-
1.0	6250	-	-	-	0.839

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

bibtex

@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

bibtex

@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}
}

	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