You can deploy this model and start earning money today!

BorisTM/
static_rus_bak

sentence-transformers

generated_from_trainer

10246566

MultipleNegativesSymmetricRankingLoss

AigizK/bashkir-russian-parallel-corpora

apache-2.0

endpoints_compatible

Views

No views yet

SentenceTransformer

This is a sentence-transformers model. It maps sentences & paragraphs to a 256-dimensional dense vector space and can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

[!IMPORTANT]
It is a static embedding model! The main purpose of it is to calculate similarity between russian and bashkir sentences.

Model Details

Model Description

Model Type: Sentence Transformer

Maximum Sequence Length: inf tokens
Output Dimensionality: 256 dimensions
Similarity Function: Cosine Similarity

Language: Bashkir

Model Sources

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

Full Model Architecture

SentenceTransformer(
  (0): StaticEmbedding(
    (embedding): EmbeddingBag(120138, 256, mode='mean')
  )
)

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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("BorisTM/static_rus_bak")
# Run inference
sentences = [
    '19.Башҡортостан Республикаһында ниндәй милләттәр йәшәй?',
    'Какие экономические реформы были проведены в России в XIX веке?',
    'Валерий Газзаев также отметил, что сегодня футбол не просто является спортом.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 256]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[ 1.0000,  0.4605, -0.0718],
#         [ 0.4605,  1.0000, -0.1179],
#         [-0.0718, -0.1179,  1.0000]])

Training Details

Training Dataset

Unnamed Dataset

Size: 10,246,566 training samples
Columns: bak and rus
Approximate statistics based on the first 1000 samples:
bak rus
type string string
details
min: 2 characters
mean: 84.01 characters
max: 536 characters
min: 2 characters
mean: 83.26 characters
max: 563 characters

Samples:

bak	rus
`Ref-de Профиль на transfermarkt.de (нем.)`	`Профиль на transfermarkt.de (нем.)`
`Уның тәүәккәл эш итеүе арҡаһында был әҙәм зарарһыҙландырыла.`	`Со свойственным ему упрямством этот человек пытается исполнить свою угрозу.`
`Ростов стадионы архитектура үҙенсәлектәре башҡа стадиондарҙыҡынан айырылып торасаҡ.`	`Нарушение технологического процесса в одном, безусловно, скажется на других этапах.`

Loss: MultipleNegativesSymmetricRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim", 4 "gather_across_devices": false 5}

Evaluation Dataset

rus_bak_real

Dataset: rus_bak_real
Size: 10,000 evaluation samples
Columns: bak and rus
Approximate statistics based on the first 1000 samples:
bak rus
type string string
details
min: 3 characters
mean: 85.78 characters
max: 1025 characters
min: 4 characters
mean: 85.84 characters
max: 967 characters

Samples:

bak	rus
`с йылдырымом юсуп а не валерой!`	`Освежает потрясающе!`
`Беренсе шуныһы ташлана күҙе - блузка уңайлы, туника, салбар.`	`Первое, что бросается в глаза - чехол, плотный и удобный.`
`Һауаның уртаса йыллыҡ театраһы — 2,3°С, ғин. уртаса температура — -15°С, июлдә — 21°С. Абсолютная максимальная температура — 38°С, абс. миним. театра — -48,0°С. Яуым-төшөмдөң уртаса йыллыҡ миҡдары — 450 мм.`	`Среднегодовая температура воздуха – 2,3°С, средняя температура янв. – -15°С, июля – 21°С. Абсолютная максимальная температура – 38°С, абс. миним. театра – -48,0°С. Среднегодовое количество осадков – 450 мм.`

Loss: MultipleNegativesSymmetricRankingLoss with these parameters:
json
1{ 2 "scale": 20.0, 3 "similarity_fct": "cos_sim", 4 "gather_across_devices": false 5}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 8192
per_device_eval_batch_size: 256
learning_rate: 0.02
weight_decay: 0.01
max_steps: 1200
warmup_ratio: 0.05
bf16: True
bf16_full_eval: True
remove_unused_columns: False
load_best_model_at_end: True

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 8192
per_device_eval_batch_size: 256
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: 0.02
weight_decay: 0.01
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
max_grad_norm: 1.0
num_train_epochs: 3.0
max_steps: 1200
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.05
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: True
fp16: False
fp16_opt_level: O1
half_precision_backend: auto
bf16_full_eval: True
fp16_full_eval: False
tf32: None
local_rank: 0
ddp_backend: None
tpu_num_cores: None
tpu_metrics_debug: False
debug: []
dataloader_drop_last: True
dataloader_num_workers: 0
dataloader_prefetch_factor: None
past_index: -1
disable_tqdm: False
remove_unused_columns: False
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}
parallelism_config: None
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch
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: False
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: 3.10.14
Sentence Transformers: 5.1.2
Transformers: 4.57.3
PyTorch: 2.3.1
Accelerate: 1.12.0
Datasets: 4.4.1
Tokenizers: 0.22.1

Model Card Authors

Malashenko Boris

Model Card Contact

quelquemath@gmail.com

	bak	rus
type	string	string
details	min: 2 characters mean: 84.01 characters max: 536 characters	min: 2 characters mean: 83.26 characters max: 563 characters