SentenceTransformer based on VoVanPhuc/sup-SimCSE-VietNamese-phobert-base

This is a sentence-transformers model finetuned from VoVanPhuc/sup-SimCSE-VietNamese-phobert-base on the json 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: VoVanPhuc/sup-SimCSE-VietNamese-phobert-base
  • Maximum Sequence Length: 256 tokens
  • Output Dimensionality: 768 dimensions
  • Similarity Function: Cosine Similarity
  • Training Dataset:
    • json

Model Sources

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 256, 'do_lower_case': False}) with Transformer model: RobertaModel 
  (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("zxcvo/sup-SimCSE-VietNamese-phobert-base-soc")
# Run inference
sentences = [
    'Vận hành một chương trình bảo mật yêu cầu các công cụ hỗ trợ kiểm soát thay đổi và theo dõi tài sản dựa trên khung phân loại tài sản.',
    'Điều gì là quan trọng nhất khi vận hành một chương trình bảo mật trong một tổ chức?',
    'Vì sao Bảo mật Điểm cuối quan trọng?',
]
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

Metric dim_768 dim_512 dim_256 dim_128 dim_64
cosine_accuracy@1 0.375 0.4062 0.3438 0.3125 0.2188
cosine_accuracy@3 0.5312 0.5 0.5312 0.5 0.4688
cosine_accuracy@5 0.5938 0.625 0.625 0.5938 0.5625
cosine_accuracy@10 0.8438 0.8438 0.7812 0.8125 0.7188
cosine_precision@1 0.375 0.4062 0.3438 0.3125 0.2188
cosine_precision@3 0.1771 0.1667 0.1771 0.1667 0.1562
cosine_precision@5 0.1187 0.125 0.125 0.1187 0.1125
cosine_precision@10 0.0844 0.0844 0.0781 0.0813 0.0719
cosine_recall@1 0.375 0.4062 0.3438 0.3125 0.2188
cosine_recall@3 0.5312 0.5 0.5312 0.5 0.4688
cosine_recall@5 0.5938 0.625 0.625 0.5938 0.5625
cosine_recall@10 0.8438 0.8438 0.7812 0.8125 0.7188
cosine_ndcg@10 0.5757 0.5859 0.5435 0.5297 0.4469
cosine_mrr@10 0.4942 0.5082 0.4699 0.4441 0.3627
cosine_map@100 0.5025 0.5154 0.4794 0.449 0.3716

Training Details

Training Dataset

json

  • Dataset: json
  • Size: 288 training samples
  • Columns: positive and anchor
  • Approximate statistics based on the first 288 samples:
    positive anchor
    type string string
    details
    • min: 14 tokens
    • mean: 49.27 tokens
    • max: 73 tokens
    • min: 9 tokens
    • mean: 20.08 tokens
    • max: 33 tokens
  • Samples:
    positive anchor
    Tường lửa bảo vệ hệ thống mạng bằng cách kiểm soát và lọc lưu lượng truy cập, ngăn chặn truy cập từ các nguồn không tin cậy và bảo vệ hệ thống khỏi các cuộc tấn công từ bên ngoài. Tường lửa bảo vệ hệ thống mạng như thế nào?
    Giám sát mạng giúp bảo vệ hệ thống CNTT bằng cách theo dõi lưu lượng truy cập, phát hiện các hành vi bất thường, và ngăn chặn các cuộc tấn công trước khi chúng có thể gây hại. Giám sát mạng có vai trò gì trong bảo vệ hệ thống CNTT?
    SIEM giúp cải thiện an ninh mạng của tổ chức bằng cách thu thập, phân tích và tương quan các sự kiện bảo mật từ nhiều nguồn khác nhau, từ đó phát hiện và cảnh báo kịp thời về các mối đe dọa, hỗ trợ xử lý sự cố nhanh chóng và hiệu quả. SIEM giúp cải thiện an ninh mạng của tổ chức như thế nào?
  • Loss: MatryoshkaLoss with these parameters:
    {
        "loss": "MultipleNegativesRankingLoss",
        "matryoshka_dims": [
            768,
            512,
            256,
            128,
            64
        ],
        "matryoshka_weights": [
            1,
            1,
            1,
            1,
            1
        ],
        "n_dims_per_step": -1
    }
    

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: epoch
  • per_device_train_batch_size: 32
  • gradient_accumulation_steps: 16
  • learning_rate: 2e-05
  • num_train_epochs: 4
  • bf16: True
  • load_best_model_at_end: True

All Hyperparameters

Click to expand
  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: epoch
  • prediction_loss_only: True
  • per_device_train_batch_size: 32
  • per_device_eval_batch_size: 8
  • per_gpu_train_batch_size: None
  • per_gpu_eval_batch_size: None
  • gradient_accumulation_steps: 16
  • eval_accumulation_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: 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: 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: True
  • 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
  • prompts: None
  • batch_sampler: batch_sampler
  • multi_dataset_batch_sampler: proportional

Training Logs

Epoch Step dim_768_cosine_ndcg@10 dim_512_cosine_ndcg@10 dim_256_cosine_ndcg@10 dim_128_cosine_ndcg@10 dim_64_cosine_ndcg@10
1.0 1 0.5303 0.5220 0.4952 0.4562 0.3810
2.0 3 0.5724 0.5737 0.5431 0.5142 0.4448
3.0 4 0.5757 0.5859 0.5435 0.5297 0.4469
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.11.9
  • Sentence Transformers: 3.3.1
  • Transformers: 4.41.2
  • PyTorch: 2.5.1+cu124
  • Accelerate: 1.1.1
  • Datasets: 2.19.1
  • 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",
}

MatryoshkaLoss

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

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}
}
Downloads last month
9
Safetensors
Model size
135M params
Tensor type
F32
·
Inference Examples
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for zxcvo/sup-SimCSE-VietNamese-phobert-base-soc

Finetuned
(3)
this model

Evaluation results