Tibetan–English Cross-Lingual Sentence Similarity

This model scores the semantic similarity between a Tibetan sentence and an English sentence. Give it any Tibetan passage and any English passage, and it returns a number between –1 and 1: values near 1 mean the two passages express the same idea; values near 0 mean they are unrelated; negative values indicate contrasting meanings.

The model was developed to support work with Tibetan Buddhist literature. As large volumes of Tibetan text are digitized and translated into English, there is a practical need to align passages across parallel texts.

It is useful for researchers, translators, and Buddhist scholars working with Tibetan–English text pairs, as well as developers building text-alignment tools.

Intended Use

The primary use case is to score the similarity between a Tibetan input and an English input:

from sentence_transformers import SentenceTransformer

model = SentenceTransformer("khyentsevision/minilm-bo-en-sim")

tibetan = "ཆོས་ཉིད་རྣམ་པར་མ་རྟོག་པའི་ཡེ་ཤེས་སུ་འདོད་པ་ལ།"
english = "This is the meaning of perceiving the true reality."

embeddings = model.encode([tibetan, english])
similarity = model.similarity(embeddings[0], embeddings[1])
print(f"Similarity score: {similarity.item():.4f}")
# ~0.71 → strong semantic match
# ~0.0  → unrelated
# negative → contrasting meanings

To rank multiple English candidates against a single Tibetan source (e.g. to compare translation options):

candidates = [
    "This is the meaning of perceiving the true reality.",
    "Make the Secret Mantra teachings prosper and flourish!",
]

source_emb = model.encode([tibetan])
cand_embs  = model.encode(candidates)
scores = model.similarity(source_emb, cand_embs)
print(scores)  # higher score = better semantic match

Model Details

Model Description

• Model Type: Sentence Transformer
• Base model: sentence-transformers/all-MiniLM-L6-v2
• Architecture: BERT-based transformer encoder with mean pooling
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 384 dimensions
• Similarity Function: Cosine Similarity

Model Sources

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

Evaluation

Metrics

Semantic Similarity

Evaluated on held-out Tibetan–English pairs spanning high, low, and near-zero similarity ranges.

Training Details

Training Dataset

• Size: 3,400,000 training samples
• Columns: Tibetan (bo), English (en), and similarity score derived from monolingual comparisons
• Loss: CosineSimilarityLoss with MSELoss

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• learning_rate: 2e-05
• warmup_ratio: 0.1
• fp16: True
• auto_find_batch_size: True

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 8
• per_device_eval_batch_size: 8
• 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: 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: 3
• 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
• 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}
• parallelism_config: None
• deepspeed: None
• label_smoothing_factor: 0.0
• optim: adamw_torch_fused
• optim_args: None
• adafactor: False
• 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
• 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: True
• 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: 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
• prompts: None
• batch_sampler: batch_sampler
• multi_dataset_batch_sampler: proportional
• router_mapping: {}
• learning_rate_mapping: {}

Framework Versions

• Python: 3.11.9
• Sentence Transformers: 5.2.0
• Transformers: 4.57.3
• PyTorch: 2.9.1+cu128
• Accelerate: 1.12.0
• Datasets: 4.4.2
• Tokenizers: 0.22.1

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",
}

Model Card Contact

jake@khyentsevision.org