This is a sentence-transformers model finetuned from BAAI/bge-base-en. 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: BAAI/bge-base-en
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 dimensions
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': 512, 'do_lower_case': True}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("mavihsrr/bge-final-small-retail-v2")
# Run inference
sentences = [
    'Extra Virgin Coconut Oil. Description :This cold pressed, pure, natural, extra virgin coconut oil because of its high saturated fat content, it is slow to oxidize and, thus, resistant to rancidification, lasting up to six months at 24  DegreeC without spoiling. This is the purest form of coconut oil, which retains all of its goodness.!',
    'Rubber Gloves - Cotton Lined, Soft & Non Slip, Medium. Description :The Super Strong Elbow Grease Rubber Gloves in large protect the hands from bacteria and chemicals during cleaning tasks, ideal for dishwashing, scrubbing task and using harmful chemicals. These high-quality designs are cotton lined, soft & non-slip gloves for ease to use. The Elbow Grease Rubber Gloves are the only gloves you will ever need. Great for domestic or commercial cleaning purpose.!',
    'Wax Candles - Metal, Smokeless, White, CD 05. Description :Enrich the ambience of the place as you place these captivating looking tealight candles. These are white in colour and round in shape. They are filled with wax and wick inside. It is suitable for decorating the house during the festive occasions and parties. These are smokeless candles that do not leave any soot residue behind. Also, the tealight candles burn fully without damaging your furniture or floor. Also, it has 25 pieces.!',
]
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: 174,064 training samples
Columns: sentence1, sentence2, and score
Approximate statistics based on the first 1000 samples:
sentence1 sentence2 score
type string string float
details
min: 12 tokens
mean: 116.54 tokens
max: 512 tokens
min: 8 tokens
mean: 111.22 tokens
max: 512 tokens
min: 0.1
mean: 0.64
max: 0.97

Samples:

sentence1	sentence2	score
Oil Clear Mud Face Pack. Description :Himalaya Oil Clear Mud Face Pack Rejuvenate your dead skin with Himalaya Oil Clear Mud Face Pack. This herbal formulation deep cleanses facial skin and clears clogged pores by absorbing excess oil and removing impurities. It helps maintain the natural pH of the skin and has deep cleansing and detoxifying properties, leaving the skin cleansed and revitalized. Fullers Earth removes deep-seated dirt and pollutants. It absorbs oil, clears clogged pores and blemishes and helps remove dead skin. Fullers Earth also helps lighten tanned skin caused by UV rays.!	Pure White Mineral Clay Anti Pollution Purity Face Wash Foam. Description :Giving your skin an oil overhaul doesn't have to be a drag. Oil stuck in your pores is what makes your skin feel oily again after a wash! POND'S Clay Foam and Mask is the most fun way to say goodbye oil and hello to an all-day matte glow. Made with 100% natural Moroccan clay that has 4x oil absorption power, it sucks out dirt and oil stuck deep within your pores. What's left behind? Skin that's glowing and matte all day long! Pond's Clay Foam is the most enjoyable and effective way to keep your skin oil-free for longer. Revolutionise face washing with the enriching power of Mineral Clay. One of the most efficacious ingredients in deep cleansing. Its enriched with skin-loving minerals to give you a bouncy glow. So, step up your deep cleansing regimen for an oil-free glow. The clay range comes in two exciting formats. The Pond's white beauty mineral clay foam brightens and smoothens your skin for an oil-free glow!...	`0.9511584211850151`
`Essence - Butter Scotch. Description :Concentrate Butterscotch Essence For Sauces, Desserts, Baking And Cakes.Butterscotch Adds A Luscious Flavor Note To Mochas, Lattes And Other Hot, Frozen And Chilled Drinks.!`	`product[object Object]Icing Sugar Icing Sugar. Description :Icing Sugar is finel...[object Object]Icing Sugar Icing Sugar. Description :This finely granulat...[object Object]Name: combined, dtype: object`	`0.9643093974992689`
`Marie Light Biscuit - Vita Orange. Description :Sunfeast Marie Light orange offer crisp & light biscuits completed with the choicest golden grains of sun-ripened oranges and wheat. It presents the only Marie biscuit in India with a stimulating, delicious orange flavour. Whats more, there is 0% transfat and 0% cholesterol making it an appetisingly vigorous biscuit.!`	Premium Wafer Bites - Dark Choco 100 g + Strawberry 100 g + Tiramisu 100 g. Description :Tasties brings you the Delicious Creamy & Crunchy Wafer Bites. Indulge in the taste of 5 wafers and 4 cream layered mini wafer bites with mouth-melting dark chocolate filling.[object Object]Tasties brings you the Delicious Creamy & Crunchy Wafer Bites. Indulge in the taste of 5 wafers and 4 cream layered mini wafer bites with mouth-melting strawberry filling.[object Object]Tasties brings you the Delicious Creamy & Crunchy Wafer Bites. Indulge in the taste of 5 wafers and 4 cream layered mini wafer bites with mouth Tiramisu hazelnut filling.[object Object][object Object]Munch on this and say bye to your small hunger pangs.!	`0.8838966912863657`

Loss: CosineSimilarityLoss with these parameters:
json
1{ 2 "loss_fct": "torch.nn.modules.loss.MSELoss" 3}

Evaluation Dataset

Unnamed Dataset

Size: 21,759 evaluation samples
Columns: sentence1, sentence2, and score
Approximate statistics based on the first 1000 samples:
sentence1 sentence2 score
type string string float
details
min: 10 tokens
mean: 121.44 tokens
max: 512 tokens
min: 10 tokens
mean: 112.13 tokens
max: 512 tokens
min: 0.1
mean: 0.61
max: 0.97

Samples:

sentence1	sentence2	score
Rose Glycerin Soap For Clean & Refreshed Skin -Cold Processed, 100% Natural & Organic. Description :Feel fresh, clean and refreshed with Rose which will leave your skin delicately scented with an uplifting rose fragrance. This soap does not use any animal product like milk or honey. It is completely vegan. Rose helps to improve the skin's appearance and to perfume the skin. It contains glycerin that softens the skin. It has no added preservatives and SLS and is 100% natural herbs and essential oils. It is a vegan product and is SLS free.!	`product[object Object]Relax Moisturising Hand Wash - Lavender & Ylang-Ylang Relax Moisturising Hand Wash - Lavender & Ylan...[object Object]Relax Moisturising Hand Wash - Lavender & Ylang-Ylang Relax Moisturising Hand Wash - Lavender & Ylan...[object Object]Name: combined, dtype: object`	`0.9641479761938232`
Dog Food - Focus Starter, Super Premium. Description :The Drools Focus, Super premium all breed formula for Puppies is formulated with the finest natural ingredients to help your dog live a long and healthy life. The result of exhaustive scientific research carried out over the years, by some of the most experienced veterinarians and nutritionists. Just like the rest of the Drools products, this one too is manufactured with a keen eye for detail and utmost care at Asias largest and most modern plant.!	Erina - Coat Cleanser. Description :Action : Dandruff control : Erina prevents the formation of dandruff on your pets skin and hair coat. Antimicrobial : Its antiseptic and antibacterial cleansing eliminates germs and improves overall skin hygiene. Erina protects the body against commonly found pathogens that cause itching and bacterial infections. Deodorant : Erinas deodorizing properties eliminate foul odor. Indications : For controlling dandruff in the hair coat. Prevention and management of pruritus (itching) and pyoder(superficial bacterial infection). Used in routine bathing as a cleanser to maintain a healthy coat.!	`0.9112330093194662`
`Fruit & Food Nibbler With Silicone Sack - Green. Description :Introducing new foods to your babys diet can be a fun learning experience as it provides him or her with new varying tastes and flavours. With Mee Mee fruit and food nibbler, your child can safely enjoy fruit and other kinds of whole foods, without the risk of choking or hurting his or her mouth.!`	Trendy Stainless Steel Bottle With Sipper Cap - Steel Matt Finish, PXP 1002 DQ. Description :Now free your environment, and yourself from the unhealthy plastic bottles and get a healthier one-time product for all your needs. These high-grade stainless steel bottles are here to enhance your dining and travelling experience, saving you from the negative effects of plastic. The single-walled steel bottles are perfect add-ons to your kitchen collection if you are looking for light-weighed, durable, classy looking product. The bottle comes with sipper & wide mouth steel cap, catering to double usage. Be it going to the gym, or sending it with the kids to school, the colourful sipper can always make it a very convenient, handy and more importantly a style-statement product. You can take it to the office or just keep on the dinner table. Open the wide mouth lid and use at ease. The bottles come with the major USP of inter-changeable lid facility. Now you can make the same steel cap bottle as ...	`0.14806349984585232`

Loss: CosineSimilarityLoss with these parameters:
json
1{ 2 "loss_fct": "torch.nn.modules.loss.MSELoss" 3}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 64
per_device_eval_batch_size: 64
learning_rate: 2e-05
warmup_ratio: 0.1
bf16: True

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 64
per_device_eval_batch_size: 64
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: 2e-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: 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: None
hub_always_push: False
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_inputs_for_metrics: False
include_for_metrics: []
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
average_tokens_across_devices: False
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.0460	500	0.1008	-
0.0092	100	0.0515	0.0453
0.0184	200	0.0532	-
0.0368	100	0.0491	0.0393
0.0735	200	0.0427	0.0333
0.1103	300	0.0373	0.0257
0.1471	400	0.0294	0.0188
0.1838	500	0.0212	0.0169
0.2206	600	0.0174	0.0131
0.2574	700	0.0145	0.0123
0.2941	800	0.0125	0.0094
0.3309	900	0.0109	0.0103
0.3676	1000	0.0102	0.0086
0.4044	1100	0.0075	0.0088
0.4412	1200	0.0077	0.0076
0.4779	1300	0.0071	0.0070
0.5147	1400	0.007	0.0072
0.5515	1500	0.0065	0.0068
0.5882	1600	0.0058	0.0073
0.625	1700	0.0064	0.0075
0.6618	1800	0.0057	0.0062
0.6985	1900	0.0055	0.0060
0.7353	2000	0.0054	0.0071
0.7721	2100	0.0055	0.0062
0.8088	2200	0.005	0.0065
0.8456	2300	0.0064	0.0061
0.8824	2400	0.0046	0.0056
0.9191	2500	0.0045	0.0051
0.9559	2600	0.0042	0.0051
0.9926	2700	0.0046	0.0055
1.0294	2800	0.0041	0.0053
1.0662	2900	0.005	0.0057
1.1029	3000	0.0033	0.0055
1.1397	3100	0.0037	0.0054
1.1765	3200	0.004	0.0052
1.2132	3300	0.0038	0.0049
1.25	3400	0.0038	0.0047
1.2868	3500	0.0035	0.0052
1.3235	3600	0.0034	0.0048
1.3603	3700	0.0035	0.0049
1.3971	3800	0.0034	0.0045
1.4338	3900	0.0037	0.0048
1.4706	4000	0.0036	0.0047
1.5074	4100	0.0031	0.0046
1.5441	4200	0.0039	0.0045
1.5809	4300	0.0033	0.0046
1.6176	4400	0.0033	0.0047
1.6544	4500	0.0035	0.0047
1.6912	4600	0.0029	0.0047
1.7279	4700	0.0035	0.0046
1.7647	4800	0.0033	0.0046
1.8015	4900	0.003	0.0046
1.8382	5000	0.0027	0.0045
1.875	5100	0.003	0.0043
1.9118	5200	0.0031	0.0046
1.9485	5300	0.0029	0.0045
1.9853	5400	0.003	0.0044
2.0221	5500	0.0031	0.0044
2.0588	5600	0.0028	0.0044
2.0956	5700	0.0032	0.0044
2.1324	5800	0.0027	0.0043
2.1691	5900	0.0032	0.0043
2.2059	6000	0.0029	0.0043
2.2426	6100	0.0028	0.0043
2.2794	6200	0.0028	0.0045
2.3162	6300	0.0032	0.0043
2.3529	6400	0.0026	0.0043
2.3897	6500	0.0026	0.0043
2.4265	6600	0.0024	0.0044
2.4632	6700	0.0024	0.0042
2.5	6800	0.0028	0.0043
2.5368	6900	0.0026	0.0043
2.5735	7000	0.0028	0.0042
2.6103	7100	0.0024	0.0043
2.6471	7200	0.0023	0.0042
2.6838	7300	0.0027	0.0041
2.7206	7400	0.0024	0.0041
2.7574	7500	0.003	0.0041
2.7941	7600	0.003	0.0041
2.8309	7700	0.0028	0.0041
2.8676	7800	0.0029	0.0041
2.9044	7900	0.0026	0.0041
2.9412	8000	0.0022	0.0041
2.9779	8100	0.0023	0.0041

Framework Versions

Python: 3.10.12
Sentence Transformers: 3.3.1
Transformers: 4.47.1
PyTorch: 2.1.0+cu118
Accelerate: 1.2.1
Datasets: 3.2.0
Tokenizers: 0.21.0

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",
}

	sentence1	sentence2	score
type	string	string	float
details	min: 12 tokens mean: 116.54 tokens max: 512 tokens	min: 8 tokens mean: 111.22 tokens max: 512 tokens	min: 0.1 mean: 0.64 max: 0.97

Ionio-ai/retail_embedding_classifier_v1

SentenceTransformer based on BAAI/bge-base-en

Model Details

Model Description

Model Sources

Full Model Architecture

Usage

Direct Usage (Sentence Transformers)

Training Details

Training Dataset

Unnamed Dataset

Evaluation Dataset

Unnamed Dataset

Training Hyperparameters

Non-Default Hyperparameters

All Hyperparameters

Training Logs

Framework Versions

Citation

BibTeX

Sentence Transformers

Ionio-ai/
retail_embedding_classifier_v1