This is a sentence-transformers model finetuned from sentence-transformers/all-mpnet-base-v2. 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: sentence-transformers/all-mpnet-base-v2
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: MPNetModel 
  (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/mpnet_fashion_5k")
# Run inference
sentences = [
    "I'm looking for a versatile and comfortable outerwear option that can provide warmth and protection from the wind. It should be suitable for both adults and kids, easy to clean, and made from a soft and breathable material.",
    "Title: Cover Custom Western Texas Stars Washable and Reusable Warm Windproof for Women Men Boys Girls Kids Descripion: ['Our Products Are Made Of High-Quality 100% Polyester Fiber, Which Is Very Soft, Breathable And Washable.']",
    'Title: Goodthreads Men\'s Soft Cotton Long-Sleeve Pullover Hoodie T-Shirt Descripion: [\'An Amazon brand - This pullover hoodie in soft cotton features a self-tie drawstring and a kangaroo pocket. Made in our Signature Tumbled Cotton for a soft, yet sturdy, hand. We utilize a unique Heritage Wash to give our garments a custom, lived-in feel right away\'\n "Goodthreads\' collection of men\'s clothing crafted with care takes wear-everywhere apparel to the next level. Create can\'t-miss pairings with long- and short-sleeve button-down shirts in standard and slim fits, plus chino pants and shorts made from wrinkle-free non-iron fabric. With these classics-and T-shirts, polo shirts, and outerwear to round out your look-Goodthreads is your go-to for wardrobe staples with the style you want."]',
]
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: 28 tokens
mean: 47.05 tokens
max: 93 tokens
min: 22 tokens
mean: 107.27 tokens
max: 128 tokens

Samples:

sentence_0	sentence_1
`I'm looking for a bulk selection of decorative pieces that can be used to create unique jewelry and craft projects. They should come in a variety of styles, with a classic gold finish, and be suitable for personalizing items like necklaces or bracelets. These accessories should also be versatile enough for various crafting applications and make a lovely gift for special occasions.`	Title: Wholesale Bulk 50PCS Mixed Gold Charms Pendants DIY for Jewelry Making and Crafting Descripion: ['50pcs Mixed KC Gold Charms Pendants for Jewelry Making, DIY Craft Charms Bulk for Necklace Bracelet Jewelry Making Crafting'[object Object] 'Color:' 'KC Gold.' 'Size:' '0.42" - 1.15" (11 - 28 mm).'[object Object] 'Main Material:' 'Alloy, Metal.' 'Package Include:'[object Object] "50pcs gold charm. 1 pcs chamois cloth of SUNEEY.(Suitable for cleaning and polishing jewelry.) Gold charm, it's fashion, creative, full of special means, is a very useful accessory,Exquisite and classical design charms.It can be used in all kinds of decoration. A good gift for yourself or friends, or birthday gift, anniversary present. Suit for key chain, bag pendant, sweater chain pendants, ornaments, escort cards, scrapbooking, and other crafts etc. These charm can be used to create earrings, necklaces, charm bracelets, and all kinds of jewelry making and craft projects. Perfect for scrapbooking project, necklace pendant drop, jewelry making accessories. Jewelry Making Accessory Mixed wholesale metal charms Assorted themes great array of subjects and styles for all kinds of interests Great for parties or groups. This gold pendant set is very charming with the unique design. It is simple and understated but gorgeous and classy. This pretty design can highlight your appearance, grasp everyone's eyes in the crowd. We’re confident that you will love them,as it will make you stand out."]
`I'm looking for a stylish and comfortable cropped top that I can easily throw on after my workouts. It should be versatile enough for casual outings and have a fit that flatters the figure.`	`Title: Core 10 Women's Soft Workout Cropped Hoodie Sweatshirt Descripion: ['An Amazon brand - This cropped Hoodie is a wardrobe-essential sweatshirt that features an easy, flattering fit for all your post-workout, layering, or everyday styling needs'[object Object] 'Empowering women to reach their full potential is at the heart of what we do. Because when you’re wearing Core 10, you’re ready to experience more. Be more. Live more.']`
`I'm looking for a comfortable and stylish top for women that has a relaxed fit and a hood. It should come in trendy colors and allow for easy movement, making it perfect for casual outings or workouts.`	`Title: PUMA Ladies' Hooded Tee Descripion: ['Features: PUMA Colors: Black, Pink, and Purple Dropped shoulder seams Side slits for enhanced range of motion Mesh interior lined hood Puma Branded taping inside of neckline Content: 60% Cotton`

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: 28 tokens mean: 47.05 tokens max: 93 tokens	min: 22 tokens mean: 107.27 tokens max: 128 tokens