carnival13/all-mpnet-base-v2-modulepred

SentenceTransformer based on sentence-transformers/all-mpnet-base-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-mpnet-base-v2 on the csv dataset. 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: 384 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Training Dataset:
  • csv

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': 384, '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})
  (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("carnival13/all-mpnet-base-v2-modulepred")
# Run inference
sentences = [
    'module: bread ambient group: bread ambient supergroup: food ambient example descriptions: 1 war 3 toastie 400 g cc 90 varburtons bread tovis snelwrspmpkin 800 g warbutons medium bread spk giant crumpets z hovis med wht 600 g sandwich thins 5 pk warb pk crumpets mission plain tortilla 25 cm warburtons 4 protein thin bagels hovis soft wet med hovis wholemefl pataks pappadums 6 pk warb so bth disc pappajuns',
    'retailer: crispcorner description: kingsmill 5050 medius bread 800 g',
    'retailer: vitalveg description: ready to eat prun',
]
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]

Evaluation

Metrics

Information Retrieval

• Dataset: sentence-transformers/all-mpnet-base-v2
• Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.4988
cosine_accuracy@3	0.6342
cosine_accuracy@5	0.7102
cosine_accuracy@10	0.7838
cosine_precision@1	0.4988
cosine_precision@3	0.2114
cosine_precision@5	0.142
cosine_precision@10	0.0784
cosine_recall@1	0.4988
cosine_recall@3	0.6342
cosine_recall@5	0.7102
cosine_recall@10	0.7838
cosine_ndcg@10	0.6324
cosine_mrr@10	0.585
cosine_map@100	0.591

Training Details

Training Dataset

csv

• Dataset: csv
• Size: 505,654 training samples
• Columns: query and full_doc
• Approximate statistics based on the first 1000 samples:

  	query	full_doc
  type	string	string
  details	min: 10 tokens mean: 14.8 tokens max: 23 tokens	min: 83 tokens mean: 115.71 tokens max: 176 tokens

• Samples:

  query	full_doc
  retailer: vitalveg description: twin xira	module: chocolate single variety group: chocolate chocolate substitutes supergroup: biscuits & confectionery & snacks example descriptions: milky way twin 43 crml prtzlarum rai galaxy mnstr pipnut 34 g dark pb cup nest mnch foge p nestle smarties shar dark choc chun x 10 pk kinder bueno 1 dr oetker 72 da poppets choc offee pouch yorkie biscuit zpk haltesers truffles bog cadbury mini snowballs p terrys choc orange 3435 g galaxy fusion dark 704 100 g
  retailer: freshnosh description: mab pop sockt	module: clothing & personal accessories group: clothing & personal accessories supergroup: clothing & personal accessories example descriptions: pk blue trad ging 40 d 3 pk opaque tight t 74 green cali jogger ss animal swing yb denim stripe pump aw 21 ff vest aw 21 girls 5 pk lounge toplo sku 1 pk fleecy tight knitted pom hat pk briefs timeless double pom pomkids hat cute face twosie sku coral jersey str pun faded petrol t 32 seamfree waist c
  retailer: nourify description: bts prwn ckt swch	module: bread sandwiches filled rolls wraps group: bread fresh fixed weight supergroup: food perishable example descriptions: us chicken may hamche sw jo dbs allbtr pp st 4 js baconfree ran posh cheesy bea naturify cb swich sp eggcress f cpdfeggbacon js cheeseonion sv duck wrap reduced price takeout egg mayo sandwich 7 takeout cheeseonion s wich 2 ad leicester plough bts cheese pman 2 1 cp bacon chese s

• Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 4
• per_device_eval_batch_size: 16
• learning_rate: 2e-05
• num_train_epochs: 1
• warmup_ratio: 0.1
• fp16: True
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 4
• 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
• 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: 1
• 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: False
• fp16: True
• 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
• eval_use_gather_object: False
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	sentence-transformers/all-mpnet-base-v2_cosine_map@100
0.0016	100	1.6195	0.2567
0.0032	200	1.47	0.3166
0.0047	300	1.2703	0.3814
0.0063	400	1.1335	0.4495
0.0079	500	0.9942	0.4827
0.0095	600	0.9004	0.5058
0.0111	700	0.8838	0.5069
0.0016	100	0.951	0.5197
0.0032	200	0.9597	0.5323
0.0047	300	0.9241	0.5406
0.0063	400	0.8225	0.5484
0.0079	500	0.7961	0.5568
0.0095	600	0.7536	0.5621
0.0111	700	0.7387	0.5623
0.0127	800	0.7716	0.5746
0.0142	900	0.7921	0.5651
0.0158	1000	0.7744	0.5707
0.0174	1100	0.8021	0.5770
0.0190	1200	0.732	0.5756
0.0206	1300	0.764	0.5798
0.0221	1400	0.7726	0.5873
0.0237	1500	0.6676	0.5921
0.0253	1600	0.6851	0.5841
0.0269	1700	0.7404	0.5964
0.0285	1800	0.6798	0.5928
0.0301	1900	0.6485	0.5753
0.0316	2000	0.649	0.5839
0.0332	2100	0.6739	0.5891
0.0348	2200	0.6616	0.6045
0.0364	2300	0.6287	0.5863
0.0380	2400	0.6602	0.5898
0.0396	2500	0.5667	0.5910

Framework Versions

• Python: 3.10.14
• Sentence Transformers: 3.1.1
• Transformers: 4.44.2
• PyTorch: 2.4.0+cu124
• Accelerate: 0.33.0
• Datasets: 2.21.0
• 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}
}