gmunkhtur/paraphrase-mongolian-minilm-mn_v2

SentenceTransformer based on sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2

This is a sentence-transformers model finetuned from sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 on the csv dataset. 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/paraphrase-multilingual-MiniLM-L12-v2
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 384 dimensions
• 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': 128, '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})
)

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("gmunkhtur/paraphrase-mongolian-minilm-mn_v2")
# Run inference
sentences = [
    '"Сэтгүүлч анд маань хоёр дахь номоо хэвлэлтээс гаргажээ"',
    '"Л.Болормаагийн хоёр дахь ном “Завгүй” хэмээн нэрийджээ."',
    'БНХАУ-ын аж үйлдвэрлэлийн үйлдвэрлэлт буурсан.',
]
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

• Datasets: dev-t and test-t
• Evaluated with EmbeddingSimilarityEvaluator

Metric	dev-t	test-t
pearson_cosine	0.9547	0.9564
spearman_cosine	0.9538	0.9567

Training Details

Training Dataset

csv

• Dataset: csv
• Size: 77,201 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 3 tokens mean: 16.02 tokens max: 96 tokens	min: 3 tokens mean: 14.66 tokens max: 87 tokens	min: -0.14 mean: 0.63 max: 1.0

• Samples:

  sentence1	sentence2	score
  Маргааш мэдээлэл өгнө	Хэвлэлийн хурал болно.	0.5448001623153687
  Дотоод аудитын шалгалтаар 2012-2013 оны үйл ажиллагаанд 16 зөрчил илэрлээ	“Монголын Хөрөнгийн Бирж” ТӨХК-ийн Төлөөлөн удирдах зөвлөл болон Гүйцэтгэх удирдлагад 13 зөвлөмж өгөгдсөн байна.	0.4059729874134063
  "хохирогчид ажлын байраар хангагдана"	"ажил олддог болно."	0.6021140813827515

• Loss: CosineSimilarityLoss with these parameters:

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

Evaluation Dataset

csv

• Dataset: csv
• Size: 77,201 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 3 tokens mean: 16.53 tokens max: 85 tokens	min: 3 tokens mean: 14.68 tokens max: 83 tokens	min: -0.04 mean: 0.62 max: 1.0

• Samples:

  sentence1	sentence2	score
  Ченжүүд мэдээллийн сүлжээтэй лут холбогдсон байх юм	"Энд ноолуурын үнэ асуусан хэдэн нөхөд яваад байна" гээд хэлчихсэн бололтой юм	0.3234536349773407
  Хий дэлбэрэлт гарсан тухай мэдээлэл байна уу?	Мэдээлэл цуглуулж байна.	0.3009476661682129
  "Энэ нь хэн нэгнээр дамжуулж биш өөрөө сонгоно гэсэн утгатай.	Өөрөө сонгоно гэсэн утгатай."	0.770484447479248

• 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: 5
• 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: 16
• 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: 5e-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: 5
• 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: 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: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss	dev-t_spearman_cosine	test-t_spearman_cosine
0	0	-	-	1.0000	-
0.1727	500	0.0046	-	-	-
0.3454	1000	0.0054	0.0042	0.9549	-
0.5181	1500	0.0069	-	-	-
0.6908	2000	0.008	0.0067	0.9298	-
0.8636	2500	0.0076	-	-	-
1.0363	3000	0.0075	0.0065	0.9317	-
1.2090	3500	0.0069	-	-	-
1.3817	4000	0.0063	0.0063	0.9366	-
1.5544	4500	0.0055	-	-	-
1.7271	5000	0.0049	0.0057	0.9411	-
1.8998	5500	0.0045	-	-	-
2.0725	6000	0.0045	0.0056	0.9405	-
2.2453	6500	0.004	-	-	-
2.4180	7000	0.0038	0.0053	0.9432	-
2.5907	7500	0.0034	-	-	-
2.7634	8000	0.0032	0.0053	0.9448	-
2.9361	8500	0.0029	-	-	-
3.1088	9000	0.0028	0.0051	0.9459	-
3.2815	9500	0.0025	-	-	-
3.4542	10000	0.0023	0.0047	0.9498	-
3.6269	10500	0.0022	-	-	-
3.7997	11000	0.0021	0.0046	0.9510	-
3.9724	11500	0.0019	-	-	-
4.1451	12000	0.0019	0.0046	0.9525	-
4.3178	12500	0.0016	-	-	-
4.4905	13000	0.0016	0.0045	0.9528	-
4.6632	13500	0.0014	-	-	-
4.8359	14000	0.0013	0.0044	0.9538	-
5.0	14475	-	-	-	0.9567

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.3.1
• Transformers: 4.47.1
• PyTorch: 2.5.1+cu121
• 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",
}