This is a sentence-transformers model finetuned from FacebookAI/roberta-base. 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: FacebookAI/roberta-base
Maximum Sequence Length: 128 tokens
Output Dimensionality: 768 tokens
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': 128, '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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("knguyennguyen/fashion_5k")
# Run inference
sentences = [
    "I'm looking for a charming ring that embodies innocence and purity, suitable for daily wear. It should have a minimalistic design and be stackable with other rings. Durability is key, and I prefer it to be available in a unique metallic finish.",
    "Title: CozzySayido Daisy Flower Ring Bands for Woman Innocent Daisy Promise Dainty Delicate Design Minimalistic Stackable Available in Silver and Rose Gold Descripion: ['“Always have something beautiful in sight, even if it’s just a daisy in a jelly glass.”'\n '- H. Jackson, Brown Jr.'\n 'Simple, sweet, stackable everyday ring for the sweetest ones.' 'Perks'\n 'Silver or rose gold Stainless steel No corrosion No peeling No spotting or staining No green fingers Resistant to perfume, sweat, and salt water Styling versatility (stack rings, minimalistic one-piece, knuckle, toe, stopper) Medical-grade stainless steel Unbreakable Unbendable Dainty, sweet, and delicate design 30-day full refund'\n 'Sizing'\n 'Available in size 3-10 (For sizing specs and how to know your size, please consult out sizing guide in the photo panel above.)'\n 'CozzySayido'\n 'We brainstorm for our customer to get the best value product. Proudly present, a daisy flower ring. We chose stainless-steel material which is no allergies for sensitive skin, durable and tough in any conditions, no matter if you wash your hand with alcohol sanitizer or washing tons of dishes or swimming in chlorine and salt water, it never gets spotting, staining or turn your finger green or any other color. Unbreakable and unbendable no matter, how you wear it.'\n 'Meaningful design, daisy flower is the symbol of innocence, purity, true love and new beginning to make every of your day the start of something new.'\n 'Designs as well as on-trend fashion jewelry for women with minimalist, dainty, sweet, and delicate style.']",
    'Title: Chelsea FC Official Soccer Gift Mens Graphic T-Shirt Navy XXL Descripion: [\'Official CFC mens T-shirt Large club crest & text print to front Garment Size (Chest): Sm. 40"; Med. 41"; Lge. 42"; XL 44"; XXL 48"; 3XL 52" 100% cotton, top quality T-shirt Many more gift ideas for him @ FootballShopOnline\']',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Training Details

Training Dataset

Unnamed Dataset

Size: 4,693 training samples
Columns: sentence_0 and sentence_1
Approximate statistics based on the first 1000 samples:
sentence_0 sentence_1
type string string
details
min: 26 tokens
mean: 44.9 tokens
max: 87 tokens
min: 21 tokens
mean: 106.64 tokens
max: 128 tokens

Samples:

sentence_0	sentence_1
`I'm looking for a spooky mask that can enhance my costume for a Halloween event. It should have a unique design, feature some lighting effects, and come with hair, suitable for adults.`	`Title: Scary Clown Mask for Penny It Cosplay Costume Halloween Led Light Up Joker Mask with Hair Latex Horror Adult Mask Party Props Descripion: ['This clown costume mask for adults is available in a standard size that fits most adults and teens and is perfect for Halloween, themed parties, haunted houses, and more. Does not include shoes or costume. Care for this 100% latex mask with attached synthetic polyester hair']`
`I'm looking for a festive accessory set to celebrate a special birthday. It should include a stylish decorative piece that can be adjusted for comfort and a fun headpiece that adds a touch of sparkle, perfect for both intimate gatherings and lively celebrations.`	Title: Birthday Queen Sash & Rhinestone Headband Set - Silver Glitter Birthday Sash Birthday Gifts for Women Birthday Party Supplies Descripion: ['"Birthday Queen" sash & rhinestone headband set ↑ No need to keep looking...it's the ultimate birthday party gifts set! ✓ Silver glitter sash + black lettering looks great on Instagram. Make the birthday queen feel special and stand out from the crowd. ✓ It's party tested and approved...will last day into night! Not only perfect for the cozy birthday parties with family but also for the crazy night at Vegas. ✓ No size trouble + Comfortable wearing: Sash can be adjust by clip on to fit from all type body figure. Headband can sit comfortably on the head and the letters are large enough to be clearly identifiable We had so much fun designing this birthday gifts set, we hope they add just as much fun to your parties too. Get the sash and headband at the same time and be prepare for the birthday celebration!']
`I'm looking for a cozy and stylish outerwear option for the colder months, ideally with a hood and a playful design. It should be warm and plush, perfect for layering, and have a comfortable fit.`	Title: OutTop Sherpa Jacket Women Fall Winter Plush Warm Hooded Stripe Color Block Thicken Warm Fleece Coats Parka Outwear Descripion: ['Package Include:1 PC Coats'[object Object] '==========================================================================='[object Object] 'SIZE TABLE' ': International standard : 1 inch = 2.54 cm☺'[object Object] "Size:S____US:4____Bust:100cm/39.37''____Sleeve:56.5cm/22.24''____Length:88cm/34.65''"[object Object] "Size:M____US:6____Bust:105cm/41.34''____Sleeve:57cm/22.44''____Length:89cm/35.04''"[object Object] "Size:L____US:8____Bust:110cm/43.31''____Sleeve:57.5cm/22.64''____Length:90cm/35.43''"[object Object] "Size:XL____US:10____Bust:115cm/45.28''____Sleeve:58cm/22.83''____Length:91cm/35.83''"[object Object] "Size:XXL____US:12____Bust:120cm/47.24''____Sleeve:58.5cm/23.03''____Length:92cm/36.22''"[object Object] "Size:XXXL____US:14____Bust:125cm/49.21''____Sleeve:59cm/23.23''____Length:93cm/36.61''"[object Object] "Size:XXXXL____US:16____Bust:130cm/51.18''____Sleeve:59.5cm/23.43''____Length:94cm/37.01''"[object Object] "Size:XXXXXL____US:18____Bust:135cm/53.15''____Sleeve:60cm/23.62''____Length:95cm/37.40''"[object Object] '==========================================================================='[object Object] 'Any questions, please feel free to contact us.☺☺' 'Delivery:'[object Object] 'Standard express would take 7-20 days to deliver. Expedited express need 5-7 days.☺☺']

Loss: MultipleNegativesRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim" 4}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 128
per_device_eval_batch_size: 128
num_train_epochs: 5
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: no
prediction_loss_only: True
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
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
num_train_epochs: 5
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.0
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: 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
eval_on_start: False
use_liger_kernel: False
eval_use_gather_object: False
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin

Framework Versions

Python: 3.11.11
Sentence Transformers: 3.1.1
Transformers: 4.45.2
PyTorch: 2.5.1+cu121
Accelerate: 1.2.1
Datasets: 3.2.0
Tokenizers: 0.20.3

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",
}

MultipleNegativesRankingLoss

bibtex

@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}
}

	sentence_0	sentence_1
type	string	string
details	min: 26 tokens mean: 44.9 tokens max: 87 tokens	min: 21 tokens mean: 106.64 tokens max: 128 tokens