metadata

base_model: sentence-transformers/all-MiniLM-L6-v2
datasets: []
language: []
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - dot_accuracy
  - manhattan_accuracy
  - euclidean_accuracy
  - max_accuracy
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:2320
  - loss:MultipleNegativesRankingLoss
widget:
  - source_sentence: DENNIE FOSTE Men's Poly Cotton Washed Light Blue Jeans(DF-JNS-015)
    sentences:
      - https://www.amazon.in/dp/B0BZDFGSCR
      - >-
        DENNIE FOSTE presents this streachable fabric Polycotton jeans. It's
        good quality fabric would certainly make you feel good and confident
        when you wear it. Comfortable front pockets, comfortable back pockets,
        highly durable and stretchable jeans for man. Perfect for casual, beach
        parties wear high on style and quality, these stretchable jeans are as
        versatile as they are comfortable. Wear it with a casual tee for a smart
        look. Wear it casually and be at ease throughout the day or it can also
        blend to perfection on your special ocassions.
      - urbano fashion mens slim fit jeans
  - source_sentence: >-
      ZESICA Women's 2023 Summer Bohemian Solid Color Lace Trim Flowy A Line
      Beach Long Maxi Skirt with Pockets
    sentences:
      - >-
        aratlench acrylic pendant necklace earrings – long statement leaf charm
        necklace tortoise resin palm leaf earrings fashion necklaces earrings
        for women girls
      - https://www.amazon.com/dp/B09X19HV5D
      - >-
        zesica womens 2023 summer bohemian solid color lace trim flowy a line
        beach long maxi skirt with pockets
  - source_sentence: >-
      DHRUVI TRENDZ Men's Shirts || Rayon Tropical Printed Shirts for Men ||
      Summer Wear Shirt for Men || Perfect for Outing || Vacation || DateWear
      Shirt for Boys || Gift for Men
    sentences:
      - >-
        om sai latest creation shirt for men  rayon shirts for men  tropical
        leaf printed short sleeve  spread collar shirts for boy  casual beach
        wear festive shirt for men
      - https://www.amazon.in/dp/B0C18PR364
      - >-
        Men's Fashion Products Are Our partywear outfit collection for men
        includes a shirt neckline, Short-sleeves, and a button placket on the
        front. Perfect Regular Fit with Best Look. simple spread collar and soft
        felt in the fabric which makes the shirt very easy and comfortable to
        wear casually. From the newest designs and trendiest styles for men we
        are making fashionable clothing affordable. Shirts feel soft and light
        on the body. Pairing with the right colored denim we can imagine the
        outfit is best suited for dining parties and night outs. Our men's
        Tropical shirts are made of the Best fabric which is lightweight and
        breathable. Perfect for summer and hot weather keeps your body dry and
        comfortable all day. This casual summer shirts design with a Fancy
        Hawaii collar, short sleeve, botton down, Tropical print and classic
        regular fit. This beach shirts with multiple unique color and pattern,
        each of which is a unique experience, make you shine this summer.
        Perfect gift for yourself, families, or friends. Perfect for camp, sun
        beach, birthday party, vacation, bachelor party, cruise, camp, or any
        casual daily wear.
  - source_sentence: >-
      Molie Bridal Austrian Crystal Necklace and Earrings Jewelry Set Gifts fit
      with Wedding Dress
    sentences:
      - >-
        You should have this jewelry set near you all the time since it is so
        fashion and eye-catching. You can wear it and have it with you to
        support you wherever you go. Make a statement with this wonderful
        jewelry set. Molie Molie has been found for many years, referred to
        "Molie", which denotes to treat all of the world's women like an Molie
        jewelry and meet their fantasies and satisfactions. We have our own
        factory to ensure our items' plating and the strict criteria of the
        plating thickness. The physical characteristics of human require us to
        adopt a higher standard of plating process. At the same time, it create
        a good condition to reduce production cost while maintain high quality
        of our item. Moreover, We are committed to provide customers with
        competitive products and best customer services, since its inception has
        been its high quality themselves, stylish design, superb manufacturing
        process. Besides, we concentrate on improving the service based on the
        creative, showing brand attributes. All in all, we take Customers'
        satisfactions as our first priority.
      - https://www.amazon.com/dp/B071VM3BKW
      - >-
        coofandy mens short sleeve hoodie relaxed fit fashion casual sweatshirts
        lightweight hip hop streetwear t shirts
  - source_sentence: Steve Madden Clutch Crossbody
    sentences:
      - https://www.amazon.com/dp/B07VCDT9VR
      - >-
        See and BSCENE with this Clear bag. Carry it as a crossbody or clutch.
        The exterior is Clear and includes an internal pouch.
      - womens dezier mens regular shirt 6032sformal1110multicolor extra large
model-index:
  - name: SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
    results:
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: Unknown
          type: unknown
        metrics:
          - type: cosine_accuracy
            value: 1
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 1
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 1
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 1
            name: Max Accuracy

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2. 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: sentence-transformers/all-MiniLM-L6-v2
Maximum Sequence Length: 256 tokens
Output Dimensionality: 384 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': 256, 'do_lower_case': False}) with Transformer model: BertModel 
  (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})
  (2): Normalize()
)

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 = [
    'Steve Madden Clutch Crossbody',
    'See and BSCENE with this Clear bag. Carry it as a crossbody or clutch. The exterior is Clear and includes an internal pouch.',
    'https://www.amazon.com/dp/B07VCDT9VR',
]
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

Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	1.0
dot_accuracy	0.0
manhattan_accuracy	1.0
euclidean_accuracy	1.0
max_accuracy	1.0

Training Details

Training Dataset

Unnamed Dataset

Size: 2,320 training samples
Columns: anchor, positive, and negative

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative
type	string	string	string
details	min: 5 tokens mean: 21.75 tokens max: 55 tokens	min: 3 tokens mean: 59.78 tokens max: 256 tokens	min: 21 tokens mean: 23.3 tokens max: 25 tokens

Samples:

anchor	positive	negative
`Shiaili Classic Plus Size Skirts for Women Flowy Pleated Midi Length Skirt`	`shiaili classic plus size skirts for women flowy pleated midi length skirt`	`https://www.amazon.com/dp/B0BMTRJRG6`
`ANRABESS Women's Casual Long Sleeve Draped Open Front Knit Pockets Long Cardigan Jackets Sweater`	`anrabess womens casual long sleeve draped open front knit pockets long cardigan jackets sweater`	`https://www.amazon.com/dp/B0B2W6QGYB`
`RipSkirt Hawaii`	Length 2 with Pockets	Quick Wrap, Quick Dry, Travel Skirt with Side Pockets

Loss: MultipleNegativesRankingLoss with these parameters:

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

Evaluation Dataset

Unnamed Dataset

Size: 580 evaluation samples
Columns: anchor, positive, and negative

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative
type	string	string	string
details	min: 4 tokens mean: 21.92 tokens max: 60 tokens	min: 3 tokens mean: 55.98 tokens max: 256 tokens	min: 20 tokens mean: 23.37 tokens max: 25 tokens

Samples:

anchor	positive	negative
`Hotouch Lightweight Crochet Cardigan for Women Long Sleeve Open Front Knit Oversized Cardigans Sweaters`	`hotouch lightweight crochet cardigan for women long sleeve open front knit oversized cardigans sweaters`	`https://www.amazon.com/dp/B0C1FM1JDZ`
`SEIKO Men's SNK809 5 Automatic Stainless Steel Watch with Black Canvas Strap`	Black dial. Silver-tone stainless steel case with a black canvas band. Automatic movement. 30 meters / 100 feet water resistance. Fixed bezel. Tang clasp. Case size 37 mm x 11 mm. Seiko SNK809 Seiko 5 Watch.The Seiko 5 Men's Automatic Black Strap Black Dial Watch is a stylish timepiece with the convenience of automatic movement. A uniquely designed, black dial features white Arabic numbers marking the hours on an inner circle and the minutes on an outer circle, while small, bar indexes encircle the dial on an outside minute track. Silver-tone hands with luminous fill make it easy to tell time day or night, and the slim second hand is detailed with a red accent. For added convenience, a day and date display are set at three o'clock. The polished stainless steel case extends to meet the black nylon strap, which wraps comfortably around the wrist and fastens with a traditional buckle. Water resistant to 30 feet (100 meters), this high-performance watch is perfect for everyday wear.This is an automatic mechanical watch. Automatic watches do not operate on batteries, instead, they are powered automatically by the movement of the wearer’s arm. If the main spring in your automatic watch is not wound sufficiently, timekeeping may become less accurate. In order to maintain accuracy, wear the watch for 8 hours or more per day, or manually wind the main spring by turning the crown. When not in use, automatic watches may be kept charged with an automatic watch winder – a watch storage unit which may be purchased separately. From Humble beginnings, Kintaro Hattori’s Vision for Seiko has become reality. A consuming passion for excellence - imprinted in our Corporate DNA passed from generation to generation. Seiko, for 125 years committed to the art and science of time. A culture of innovation connects a 19th century Tokyo clock shop with 20th century advances in timekeeping to an extraordinary 21st century "quiet revolution." Continually driven by dedication and passion, established a multitude of world’s first technologies… transforming the principles of timekeeping. The first quartz wristwatch – changed the history of time. The first Kinetic – marked a new era in quartz watch technology. In 1969, Seiko Astron, the first quartz wristwatch - was introduced. In an instant, Seiko exponentially improved the accuracy of wristwatches –And Seiko technology firmly established today’s standard in Olympic and sports timing. 1984, another celebrated first – Kinetic Technology – powered by body movement. Kinetic – a quartz mechanism with unparalleled accuracy –the driving force behind more world’s firsts. Kinetic Chronograph – the next generation of high performance timekeeping. Kinetic Auto Relay – automatically resets to the correct time. Kinetic Perpetual - combining the date perfect technology of perpetual calendar with the genius of Kinetic Auto Relay. And now Kinetic Direct Drive – move, and the watch is powered automatically. Or hand wind it and see the power you are generating in real time. In the realm of fine watches, time is measured by Seiko innovation – A heritage of dedication to the art and science of time.See more	`https://www.amazon.com/dp/B002SSUQFG`
`Carhartt Men's Rain Defender Loose Fit Midweight Thermal-Lined Full-Zip Sweatshirt`	This men's full-zip sweatshirt is equipped for light rain. Made from midweight fleece with a water-repellent finish and thermal lining. Features inner and outer pockets that include storage for your phone. 10.5-ounce, 50% cotton / 50% polyester fleece. Polyester fleece lining for warmth. Rain Defender® durable water repellent (DWR) keeps you dry and moving in light rain. Original fit. Full-zip front with brass zipper. Attached, thermal-lined three-piece hood with drawcord closure. Spandex-reinforced rib-knit cuffs and waist help keep out the cold. Two front handwarmer pockets with flaps for added security. Hidden media pocket. Inside pocket with zipper closure. Locker loop.	`https://www.amazon.com/dp/B08BG5V4KR`

Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_eval_batch_size: 16
learning_rate: 3e-05
lr_scheduler_type: cosine
warmup_ratio: 0.1
load_best_model_at_end: True
ddp_find_unused_parameters: False

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 8
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: 3e-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: 3
max_steps: -1
lr_scheduler_type: cosine
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: 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: True
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: False
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
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	loss	max_accuracy
0.0862	25	0.3631	-	-
0.1724	50	0.1219	-	-
0.2586	75	0.1909	-	-
0.3448	100	0.24	-	-
0.4310	125	0.1607	-	-
0.5172	150	0.1103	-	-
0.6034	175	0.0952	-	-
0.6897	200	0.1139	-	-
0.7759	225	0.1335	-	-
0.8621	250	0.0758	-	-
0.9483	275	0.0902	-	-
1.0	290	-	0.0700	1.0
1.0345	300	0.0951	-	-
1.1207	325	0.0373	-	-
1.2069	350	0.086	-	-
1.2931	375	0.0418	-	-
1.3793	400	0.0522	-	-
1.4655	425	0.0387	-	-
1.5517	450	0.0217	-	-
1.6379	475	0.0455	-	-
1.7241	500	0.0424	-	-
1.8103	525	0.0238	-	-
1.8966	550	0.0355	-	-
1.9828	575	0.0283	-	-
2.0	580	-	0.0597	1.0
2.0690	600	0.0213	-	-
2.1552	625	0.0219	-	-
2.2414	650	0.0254	-	-
2.3276	675	0.0204	-	-
2.4138	700	0.0052	-	-
2.5	725	0.0248	-	-
2.5862	750	0.0507	-	-
2.6724	775	0.0191	-	-
2.7586	800	0.018	-	-
2.8448	825	0.0176	-	-
2.9310	850	0.0193	-	-
3.0	870	-	0.0566	1.0

Framework Versions

Python: 3.10.14
Sentence Transformers: 3.0.1
Transformers: 4.42.2
PyTorch: 2.3.0
Accelerate: 0.31.0
Datasets: 2.19.1
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",
}

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}
}