dawn78/minilm6_perfumerecommender_v2

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 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': 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 = [
    'lalibela by memo paris is an enchanting fragrance that transports you to a lush garden awash with the vibrancy of blooming flowers this scent harmoniously blends rich floral notes with a warm sweet undertone creating a captivating atmosphere that feels both opulent and inviting the essence of tender blooms is beautifully complemented by a hint of creamy sweetness enveloping the wearer in a soft and sensuous embrace users describe lalibela as a sophisticated and luxurious perfume often noting its unique complexity that unfolds throughout the day with a cozy warmth reminiscent of cherished memories it evokes feelings of comfort and elegance many appreciate its longlasting presence making it a favorite for both day and evening wear reviewers highlight the fragrances ability to turn heads leaving a memorable impression without being overwhelming overall lalibela is a masterful olfactory experience that embodies the spirit of femininity perfect for those seeking a signature scent infused with both richness and charm',
    'vanilla',
    'green notes',
]
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

Semantic Similarity

• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.9222
spearman_cosine	0.8309

Training Details

Training Dataset

Unnamed Dataset

• Size: 7,022 training samples
• Columns: sentence_0, sentence_1, and label
• Approximate statistics based on the first 1000 samples:

  	sentence_0	sentence_1	label
  type	string	string	float
  details	min: 12 tokens mean: 178.91 tokens max: 256 tokens	min: 3 tokens mean: 3.93 tokens max: 7 tokens	min: 0.0 mean: 0.51 max: 1.0

• Samples:

  sentence_0	sentence_1	label
  americas by mahogany is an invigorating fragrance crafted for the modern man embodying the spirit of adventure and exploration this scent opens with a refreshing burst that hints at invigorating elements balanced with a spicy warmth that captivates the senses users often describe it as a harmonious blend of fresh and woody accords creating an atmosphere that is both vibrant and grounding this perfumes unique composition has garnered praise for its longevity and versatility making it suitable for both day and evening wear reviewers highlight its ability to evoke a sense of confidence and sophistication likening its character to a breezy stroll through verdant landscapes complemented by hints of rich aromatic woods the overall mood is one of refined masculinity inviting wearers to embrace their individuality while leaving a memorable impression with a commendable rating of 426 out of 5 americas resonates with those seeking a fragrance that is fresh yet complex encapsulating the essence o...	black pepper	1.0
  rosa gallica by brecourt is an exquisite unisex fragrance that encapsulates a warm and inviting aura perfect for those who appreciate the elegance of nature intertwined with a touch of mystique with its alluring blend of floral and woody elements this fragrance wraps the wearer in a soft balsamic embrace that evokes a sense of serene sophistication crafted by renowned perfumer emilie bouge rosa gallica has garnered a favorable reception from users who describe it as both captivating and refined reviewers often highlight its unique combination of spicy warmth and delicate floral notes creating an atmosphere that is both romantic and grounded its celebrated for its longevity allowing the wearer to carry its gentle yet complex character throughout the day fans of this fragrance appreciate its versatility noting it as suitable for both day and evening wear with many finding it perfect for cooler weather the overall sentiment points to a fragrance that feels timeless and elegant evoking the...	pimento	0.0
  roberto cavalli nero assoluto is a mesmerizing fragrance that embodies a rich and sensual experience perfect for the confident modern woman with its intriguing blend of velvety florals and warm undertones this perfume radiates an aura of elegance and sophistication users frequently describe it as both alluring and comforting invoking sensations of nightfall enhanced by an embrace of soft sweetness the fragrance opens with captivating floral notes that invite curiosity and provide a fresh inviting introduction as it evolves deeper and warmer accords emerge creating a luxurious complexity that many wearers find addictive the interplay between the sweet and powdery elements gives it a creamy enveloping quality that makes it an excellent choice for evening wear or special occasions with an impressive rating from fragrance enthusiasts many reviewers appreciate its longevity and sillage stating that it lingers beautifully without being overpowering the overall impression of nero assoluto is ...	blonde woods	1.0

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• num_train_epochs: 4
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 32
• per_device_eval_batch_size: 32
• 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: 4
• 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: 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: round_robin

Training Logs

Epoch	Step	Training Loss	spearman_cosine
0.2273	50	-	0.2856
0.4545	100	-	0.7237
0.6818	150	-	0.7814
0.9091	200	-	0.8037
1.0	220	-	0.8096
1.1364	250	-	0.8067
1.3636	300	-	0.8130
1.5909	350	-	0.8195
1.8182	400	-	0.8221
2.0	440	-	0.8222
2.0455	450	-	0.8229
2.2727	500	0.0988	0.8177
2.5	550	-	0.8277
2.7273	600	-	0.8273
2.9545	650	-	0.8287
3.0	660	-	0.8287
3.1818	700	-	0.8269
3.4091	750	-	0.8295
3.6364	800	-	0.8301
3.8636	850	-	0.8306
4.0	880	-	0.8309

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 3.3.1
• Transformers: 4.47.1
• PyTorch: 2.5.1+cu124
• Accelerate: 1.2.1
• Datasets: 3.2.0
• Tokenizers: 0.21.0

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",
}