PrabalAryal/Sentence_Transformer_v0.0.1

SentenceTransformer based on sentence-transformers/paraphrase-multilingual-mpnet-base-v2

This is a sentence-transformers model finetuned from sentence-transformers/paraphrase-multilingual-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/paraphrase-multilingual-mpnet-base-v2
• Maximum Sequence Length: 64 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': 64, 'do_lower_case': False}) with Transformer model: XLMRobertaModel 
  (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.

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("PrabalAryal/Sentence_Transformer_v0.0.1")
# Run inference
sentences = [
    'De sproeier van de douche',
    'De douchesproeier is kapot',
    'De deur naar buiten',
]
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

Binary Classification

• Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.9909
cosine_accuracy_threshold	0.7341
cosine_f1	0.991
cosine_f1_threshold	0.7341
cosine_precision	0.984
cosine_recall	0.998
cosine_ap	0.9956
cosine_mcc	0.9819

Training Details

Training Dataset

Unnamed Dataset

• Size: 8,884 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: 3 tokens mean: 8.6 tokens max: 18 tokens	min: 4 tokens mean: 8.86 tokens max: 21 tokens	min: 0.0 mean: 0.51 max: 1.0

• Samples:

  sentence_0	sentence_1	label
  Het slot is kapot	Schade aan de sluiting	1.0
  Ik kan er niet uit met de auto	De uitrit is versperd	1.0
  De afvoer van de wasmachine is stuk	Lekkende kranen of leidingen	0.0

• 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: 64
• per_device_eval_batch_size: 64
• num_train_epochs: 8
• fp16: True
• 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: 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: 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: 8
• 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: 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: None
• hub_always_push: False
• hub_revision: None
• 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
• 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
• liger_kernel_config: None
• 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	cosine_ap
0.1942	27	-	0.8916
0.3885	54	-	0.9339
0.5827	81	-	0.9614
0.7770	108	-	0.9740
0.9712	135	-	0.9706
1.0	139	-	0.9732
1.1655	162	-	0.9763
1.3597	189	-	0.9831
1.5540	216	-	0.9845
1.7482	243	-	0.9858
1.9424	270	-	0.9886
2.0	278	-	0.9896
2.1367	297	-	0.9904
2.3309	324	-	0.9900
2.5252	351	-	0.9907
2.7194	378	-	0.9921
2.9137	405	-	0.9919
3.0	417	-	0.9917
3.1079	432	-	0.9933
3.3022	459	-	0.9923
3.4964	486	-	0.9911
3.5971	500	3.1664	-
3.6906	513	-	0.9936
3.8849	540	-	0.9926
4.0	556	-	0.9928
4.0791	567	-	0.9931
4.2734	594	-	0.9949
4.4676	621	-	0.9940
4.6619	648	-	0.9930
4.8561	675	-	0.9932
5.0	695	-	0.9935
5.0504	702	-	0.9938
5.2446	729	-	0.9950
5.4388	756	-	0.9949
5.6331	783	-	0.9948
5.8273	810	-	0.9948
6.0	834	-	0.9946
6.0216	837	-	0.9945
6.2158	864	-	0.9955
6.4101	891	-	0.9955
6.6043	918	-	0.9955
6.7986	945	-	0.9956

Framework Versions

• Python: 3.11.13
• Sentence Transformers: 4.1.0
• Transformers: 4.53.3
• PyTorch: 2.6.0+cu124
• Accelerate: 1.9.0
• Datasets: 4.4.1
• Tokenizers: 0.21.2

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}
}