votanthanh32004/paraphrase-multilingual-mpnet-base-v2-experience

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: 128 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': 128, 'do_lower_case': False, 'architecture': '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("sentence_transformers_model_id")
# Run inference
sentences = [
    'Yêu cầu 10+ năm kinh nghiệm, dẫn dắt mảng an toàn dược phẩm toàn cầu, am hiểu các quy định báo cáo biến cố bất lợi của FDA/EMA và dẫn dắt các buổi điều trần về an toàn thuốc.',
    'Bác sĩ điều trị tại khoa Tim mạch trong 11 năm, am hiểu về tác dụng phụ của thuốc trên bệnh nhân thực tế, nhưng chưa từng làm việc trong môi trường công ty dược hay quản trị quy trình an toàn thuốc theo chuẩn quốc tế.',
    'Developed specialized computer vision models for automated farming equipment over 5 years, expertly processing drone imagery to assess crop health indices and drastically reducing chemical waste through targeted spraying algorithms.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.4418, 0.1802],
#         [0.4418, 1.0000, 0.0758],
#         [0.1802, 0.0758, 1.0000]])

Evaluation

Metrics

Semantic Similarity

• Dataset: experience_val
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.9441
spearman_cosine	0.9217

Training Details

Training Dataset

Unnamed Dataset

• Size: 1,350 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: 10 tokens mean: 47.1 tokens max: 73 tokens	min: 20 tokens mean: 47.36 tokens max: 91 tokens	min: 0.1 mean: 0.45 max: 0.99

• Samples:

  sentence_0	sentence_1	label
  Requires 6+ years of experience acting as a formal people manager for a team of DevOps engineers, conducting 1-on-1s, managing career progression, and resolving personnel blockers.	Designed robust OCR pipelines using Tesseract over 7 years, specifically writing algorithms to automatically deskew and preprocess degraded scanned documents to improve text extraction.	0.15
  Yêu cầu 10+ năm kinh nghiệm dẫn dắt các chiến dịch di trú hệ thống (Migration) từ On-premise lên Cloud (AWS/Azure), xử lý các bài toán kỹ thuật nợ và tái cấu trúc hệ thống Legacy.	Giáo viên dạy môn Hóa học tại trường cấp 3 trong 11 năm, hướng dẫn học sinh thực hiện các phản ứng hóa học trong phòng thí nghiệm và ôn tập cho kỳ thi THPT quốc gia.	0.15
  Seeking a highly adaptable analyst with 3+ years of experience at a clean-tech startup, building financial models from scratch without existing templates, and pivoting quickly based on shifting venture capital priorities.	Worked as a senior energy analyst at a massive legacy utility provider for 15 years, strictly utilizing rigid, decade-old proprietary software templates to forecast baseline coal consumption.	0.4

• 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: 16
• per_device_eval_batch_size: 16
• num_train_epochs: 4
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• 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: None
• warmup_ratio: None
• warmup_steps: 0
• log_level: passive
• log_level_replica: warning
• log_on_each_node: True
• logging_nan_inf_filter: True
• enable_jit_checkpoint: False
• save_on_each_node: False
• save_only_model: False
• restore_callback_states_from_checkpoint: False
• use_cpu: False
• seed: 42
• data_seed: None
• bf16: False
• fp16: False
• bf16_full_eval: False
• fp16_full_eval: False
• tf32: None
• local_rank: -1
• ddp_backend: None
• debug: []
• dataloader_drop_last: False
• dataloader_num_workers: 0
• dataloader_prefetch_factor: None
• disable_tqdm: False
• remove_unused_columns: True
• label_names: None
• load_best_model_at_end: False
• ignore_data_skip: False
• fsdp: []
• fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}
• accelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• optim_args: None
• group_by_length: False
• length_column_name: length
• project: huggingface
• trackio_space_id: trackio
• 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
• 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_for_metrics: []
• eval_do_concat_batches: True
• auto_find_batch_size: False
• full_determinism: False
• ddp_timeout: 1800
• torch_compile: False
• torch_compile_backend: None
• torch_compile_mode: None
• include_num_input_tokens_seen: no
• 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: True
• use_cache: False
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: round_robin
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Epoch	Step	experience_val_spearman_cosine
1.0	85	0.9181
1.1765	100	0.9177
2.0	170	0.9217

Framework Versions

• Python: 3.12.12
• Sentence Transformers: 5.2.3
• Transformers: 5.0.0
• PyTorch: 2.10.0+cu128
• Accelerate: 1.13.0
• Datasets: 4.0.0
• Tokenizers: 0.22.2

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",
}