Training Script: github.com/adeshkin/khakas-emb/experiments/finetune_mnrl_1.py

SentenceTransformer based on sentence-transformers/LaBSE

This is a sentence-transformers model finetuned from sentence-transformers/LaBSE on the khakas-russian-parallel-corpus dataset. It maps sentences & paragraphs to a 768-dimensional dense vector space and can be used for retrieval.

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: sentence-transformers/LaBSE
• Maximum Sequence Length: 256 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity
• Supported Modality: Text
• Training Dataset:
  • khakas-russian-parallel-corpus
• Languages: kjh, ru

Model Sources

• Documentation: Sentence Transformers Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'BertModel'})
  (1): Pooling({'embedding_dimension': 768, 'pooling_mode': 'cls', 'include_prompt': True})
  (2): Dense({'in_features': 768, 'out_features': 768, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh', 'module_input_name': 'sentence_embedding', 'module_output_name': 'sentence_embedding'})
  (3): 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("adeshkin/labse-kjh-ru-mnrl-1")
# Run inference
sentences = [
    'Тӧреенде сағыпчатхан чуртас узуны ортымах чуртас узунын санирында тузаланылча.',
    'Ожидаемая продолжительность жизни при рождении используется в качестве средней продолжительности жизни.',
    'Уранча. Торгаях, ты зачем беспокоишь гостя нашими заботами?',
]
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,  0.7541, -0.0902],
#         [ 0.7541,  1.0000, -0.1342],
#         [-0.0902, -0.1342,  1.0000]])

Evaluation

Metrics

Translation

• Dataset: kjh-ru-random
• Evaluated with TranslationEvaluator

Training Details

Training Dataset

khakas-russian-parallel-corpus

• Dataset: khakas-russian-parallel-corpus at 318e0f5
• Size: 157,620 training samples
• Columns: kjh and ru
• Approximate statistics based on the first 1000 samples:

  	kjh	ru
  type	string	string
  details	min: 6 tokensmean: 28.24 tokensmax: 111 tokens	min: 4 tokensmean: 20.11 tokensmax: 84 tokens

• Samples:

  kjh	ru
  Тӧреенде сағыпчатхан чуртас узуны ортымах чуртас узунын санирында тузаланылча.	Ожидаемая продолжительность жизни при рождении используется в качестве средней продолжительности жизни.
  ТЕЛЕФОН НОМЕРІ (пар полза) 11.	НОМЕР ТЕЛЕФОНА (если имеется) 11.
  Танығлар піріктірілген полза, ол кӧзідіг прай тоғынчатхан танығларның хыраның узунының суммазынаң пиріл парған.	Если признаки составные, данный показатель представлен суммой длины поля всех задействованных признаков.

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

Evaluation Dataset

khakas-russian-parallel-corpus

• Dataset: khakas-russian-parallel-corpus at 318e0f5
• Size: 1,593 evaluation samples
• Columns: kjh and ru
• Approximate statistics based on the first 1000 samples:

  	kjh	ru
  type	string	string
  details	min: 7 tokensmean: 28.36 tokensmax: 146 tokens	min: 5 tokensmean: 20.47 tokensmax: 97 tokens

• Samples:

  kjh	ru
  Чуртас тооза пос чирінің омазын чӱреенде ал чӧрген, хайда даа полза, Хакас чирінеңер кӧглеен.	Всю жизнь носил в сердце образ своей земли, где бы ни был, воспевал Хакасскую землю.
  2 Пил палых сурча	2 Таймень-рыба спрашивает:
  чолға сығар тим чит килген	наступила пора собираться в путь

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim",
      "gather_across_devices": false,
      "directions": [
          "query_to_doc"
      ],
      "partition_mode": "joint",
      "hardness_mode": null,
      "hardness_strength": 0.0
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• learning_rate: 2e-05
• num_train_epochs: 1
• warmup_steps: 1000
• fp16: True

All Hyperparameters

• do_predict: False
• prediction_loss_only: True
• per_device_train_batch_size: 8
• per_device_eval_batch_size: 8
• 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: None
• warmup_ratio: None
• warmup_steps: 1000
• 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: True
• 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: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Training Logs

Training Time

• Training: 1.4 hours

Framework Versions

• Python: 3.12.13
• Sentence Transformers: 5.4.1
• 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",
}

MultipleNegativesRankingLoss

@misc{oord2019representationlearningcontrastivepredictive,
      title={Representation Learning with Contrastive Predictive Coding},
      author={Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      year={2019},
      eprint={1807.03748},
      archivePrefix={arXiv},
      primaryClass={cs.LG},
      url={https://arxiv.org/abs/1807.03748},
}