SentenceTransformer based on BAAI/bge-m3
This is a sentence-transformers model finetuned from BAAI/bge-m3. It maps sentences & paragraphs to a 1024-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: BAAI/bge-m3
- Maximum Sequence Length: 128 tokens
- Output Dimensionality: 1024 tokens
- Similarity Function: Cosine Similarity
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: XLMRobertaModel
(1): Pooling({'word_embedding_dimension': 1024, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False, 'include_prompt': True})
(2): 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("alperctnkaya/bge-m3-distilled-en-tr")
# Run inference
sentences = [
'Nippon Paint Garden Furniture Maintenance Oil is a high quality maintenance oil produced with a mixture of specially selected natural oils, specially developed for the care of hard woods such as teak, and can be applied to other wood types.',
'Nippon Paint Bahçe Mobilyası Bakım Yağı, özel olarak seçilen doğal yağların karışımı ile üretilen, özellikle teak gibi sert ahşapların bakımı için özel olarak geliştirilmiş, diğer ahşap türlerine de uygulanabilen üstün nitelikli bakım yağıdır.',
'Stover Gönüllü Faaliyet Ödülüne layık görülenlerin her biri, kâr amacı gütmeyen kendi seçtiği bir kuruluşa ödenmek üzere verilen 5000 Amerikan doları değerinde bir çeki içeren bir hatıra ödülünün yanı sıra, resmi bir törende genel başkan ve CEO tarafından verilen özel bir takdirnameye hak kazanacaktır.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]
# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]
Evaluation
Metrics
Knowledge Distillation
- Dataset:
eval
- Evaluated with
MSEEvaluator
Metric | Value |
---|---|
negative_mse | -0.039 |
Translation
- Dataset:
eval
- Evaluated with
TranslationEvaluator
Metric | Value |
---|---|
src2trg_accuracy | 0.8951 |
trg2src_accuracy | 0.8837 |
mean_accuracy | 0.8894 |
Training Details
Training Hyperparameters
Non-Default Hyperparameters
eval_strategy
: epochper_device_train_batch_size
: 16per_device_eval_batch_size
: 16learning_rate
: 2e-05warmup_ratio
: 0.1fp16
: True
All Hyperparameters
Click to expand
overwrite_output_dir
: Falsedo_predict
: Falseeval_strategy
: epochprediction_loss_only
: Trueper_device_train_batch_size
: 16per_device_eval_batch_size
: 16per_gpu_train_batch_size
: Noneper_gpu_eval_batch_size
: Nonegradient_accumulation_steps
: 1eval_accumulation_steps
: Nonetorch_empty_cache_steps
: Nonelearning_rate
: 2e-05weight_decay
: 0.0adam_beta1
: 0.9adam_beta2
: 0.999adam_epsilon
: 1e-08max_grad_norm
: 1.0num_train_epochs
: 3max_steps
: -1lr_scheduler_type
: linearlr_scheduler_kwargs
: {}warmup_ratio
: 0.1warmup_steps
: 0log_level
: passivelog_level_replica
: warninglog_on_each_node
: Truelogging_nan_inf_filter
: Truesave_safetensors
: Truesave_on_each_node
: Falsesave_only_model
: Falserestore_callback_states_from_checkpoint
: Falseno_cuda
: Falseuse_cpu
: Falseuse_mps_device
: Falseseed
: 42data_seed
: Nonejit_mode_eval
: Falseuse_ipex
: Falsebf16
: Falsefp16
: Truefp16_opt_level
: O1half_precision_backend
: autobf16_full_eval
: Falsefp16_full_eval
: Falsetf32
: Nonelocal_rank
: 0ddp_backend
: Nonetpu_num_cores
: Nonetpu_metrics_debug
: Falsedebug
: []dataloader_drop_last
: Falsedataloader_num_workers
: 0dataloader_prefetch_factor
: Nonepast_index
: -1disable_tqdm
: Falseremove_unused_columns
: Truelabel_names
: Noneload_best_model_at_end
: Falseignore_data_skip
: Falsefsdp
: []fsdp_min_num_params
: 0fsdp_config
: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': False}fsdp_transformer_layer_cls_to_wrap
: Noneaccelerator_config
: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}deepspeed
: Nonelabel_smoothing_factor
: 0.0optim
: adamw_torchoptim_args
: Noneadafactor
: Falsegroup_by_length
: Falselength_column_name
: lengthddp_find_unused_parameters
: Noneddp_bucket_cap_mb
: Noneddp_broadcast_buffers
: Falsedataloader_pin_memory
: Truedataloader_persistent_workers
: Falseskip_memory_metrics
: Trueuse_legacy_prediction_loop
: Falsepush_to_hub
: Falseresume_from_checkpoint
: Nonehub_model_id
: Nonehub_strategy
: every_savehub_private_repo
: Falsehub_always_push
: Falsegradient_checkpointing
: Falsegradient_checkpointing_kwargs
: Noneinclude_inputs_for_metrics
: Falseeval_do_concat_batches
: Truefp16_backend
: autopush_to_hub_model_id
: Nonepush_to_hub_organization
: Nonemp_parameters
:auto_find_batch_size
: Falsefull_determinism
: Falsetorchdynamo
: Noneray_scope
: lastddp_timeout
: 1800torch_compile
: Falsetorch_compile_backend
: Nonetorch_compile_mode
: Nonedispatch_batches
: Nonesplit_batches
: Noneinclude_tokens_per_second
: Falseinclude_num_input_tokens_seen
: Falseneftune_noise_alpha
: Noneoptim_target_modules
: Nonebatch_eval_metrics
: Falseeval_on_start
: Falseeval_use_gather_object
: Falsebatch_sampler
: batch_samplermulti_dataset_batch_sampler
: proportional
Training Logs
Click to expand
Epoch | Step | Training Loss | loss | eval_mean_accuracy | eval_negative_mse |
---|---|---|---|---|---|
0.02 | 100 | 0.0019 | - | - | - |
0.04 | 200 | 0.0013 | - | - | - |
0.06 | 300 | 0.0008 | - | - | - |
0.08 | 400 | 0.0008 | - | - | - |
0.1 | 500 | 0.0008 | - | - | - |
0.12 | 600 | 0.0007 | - | - | - |
0.14 | 700 | 0.0007 | - | - | - |
0.16 | 800 | 0.0007 | - | - | - |
0.18 | 900 | 0.0007 | - | - | - |
0.2 | 1000 | 0.0007 | - | - | - |
0.22 | 1100 | 0.0007 | - | - | - |
0.24 | 1200 | 0.0006 | - | - | - |
0.26 | 1300 | 0.0006 | - | - | - |
0.28 | 1400 | 0.0006 | - | - | - |
0.3 | 1500 | 0.0006 | - | - | - |
0.32 | 1600 | 0.0006 | - | - | - |
0.34 | 1700 | 0.0006 | - | - | - |
0.36 | 1800 | 0.0006 | - | - | - |
0.38 | 1900 | 0.0006 | - | - | - |
0.4 | 2000 | 0.0006 | - | - | - |
0.42 | 2100 | 0.0006 | - | - | - |
0.44 | 2200 | 0.0006 | - | - | - |
0.46 | 2300 | 0.0005 | - | - | - |
0.48 | 2400 | 0.0005 | - | - | - |
0.5 | 2500 | 0.0005 | - | - | - |
0.52 | 2600 | 0.0005 | - | - | - |
0.54 | 2700 | 0.0005 | - | - | - |
0.56 | 2800 | 0.0005 | - | - | - |
0.58 | 2900 | 0.0005 | - | - | - |
0.6 | 3000 | 0.0005 | - | - | - |
0.62 | 3100 | 0.0005 | - | - | - |
0.64 | 3200 | 0.0005 | - | - | - |
0.66 | 3300 | 0.0005 | - | - | - |
0.68 | 3400 | 0.0005 | - | - | - |
0.7 | 3500 | 0.0005 | - | - | - |
0.72 | 3600 | 0.0005 | - | - | - |
0.74 | 3700 | 0.0005 | - | - | - |
0.76 | 3800 | 0.0005 | - | - | - |
0.78 | 3900 | 0.0005 | - | - | - |
0.8 | 4000 | 0.0005 | - | - | - |
0.82 | 4100 | 0.0005 | - | - | - |
0.84 | 4200 | 0.0005 | - | - | - |
0.86 | 4300 | 0.0005 | - | - | - |
0.88 | 4400 | 0.0005 | - | - | - |
0.9 | 4500 | 0.0005 | - | - | - |
0.92 | 4600 | 0.0005 | - | - | - |
0.94 | 4700 | 0.0005 | - | - | - |
0.96 | 4800 | 0.0005 | - | - | - |
0.98 | 4900 | 0.0005 | - | - | - |
1.0 | 5000 | 0.0005 | 0.0004 | 0.8591 | -0.0453 |
1.02 | 5100 | 0.0005 | - | - | - |
1.04 | 5200 | 0.0005 | - | - | - |
1.06 | 5300 | 0.0004 | - | - | - |
1.08 | 5400 | 0.0004 | - | - | - |
1.1 | 5500 | 0.0004 | - | - | - |
1.12 | 5600 | 0.0004 | - | - | - |
1.1400 | 5700 | 0.0004 | - | - | - |
1.16 | 5800 | 0.0004 | - | - | - |
1.18 | 5900 | 0.0004 | - | - | - |
1.2 | 6000 | 0.0004 | - | - | - |
1.22 | 6100 | 0.0004 | - | - | - |
1.24 | 6200 | 0.0004 | - | - | - |
1.26 | 6300 | 0.0004 | - | - | - |
1.28 | 6400 | 0.0004 | - | - | - |
1.3 | 6500 | 0.0004 | - | - | - |
1.32 | 6600 | 0.0004 | - | - | - |
1.34 | 6700 | 0.0004 | - | - | - |
1.3600 | 6800 | 0.0004 | - | - | - |
1.38 | 6900 | 0.0004 | - | - | - |
1.4 | 7000 | 0.0004 | - | - | - |
1.42 | 7100 | 0.0004 | - | - | - |
1.44 | 7200 | 0.0004 | - | - | - |
1.46 | 7300 | 0.0004 | - | - | - |
1.48 | 7400 | 0.0004 | - | - | - |
1.5 | 7500 | 0.0004 | - | - | - |
1.52 | 7600 | 0.0004 | - | - | - |
1.54 | 7700 | 0.0004 | - | - | - |
1.56 | 7800 | 0.0004 | - | - | - |
1.58 | 7900 | 0.0004 | - | - | - |
1.6 | 8000 | 0.0004 | - | - | - |
1.62 | 8100 | 0.0004 | - | - | - |
1.6400 | 8200 | 0.0004 | - | - | - |
1.6600 | 8300 | 0.0004 | - | - | - |
1.6800 | 8400 | 0.0004 | - | - | - |
1.7 | 8500 | 0.0004 | - | - | - |
1.72 | 8600 | 0.0004 | - | - | - |
1.74 | 8700 | 0.0004 | - | - | - |
1.76 | 8800 | 0.0004 | - | - | - |
1.78 | 8900 | 0.0004 | - | - | - |
1.8 | 9000 | 0.0004 | - | - | - |
1.8200 | 9100 | 0.0004 | - | - | - |
1.8400 | 9200 | 0.0004 | - | - | - |
1.8600 | 9300 | 0.0004 | - | - | - |
1.88 | 9400 | 0.0004 | - | - | - |
1.9 | 9500 | 0.0004 | - | - | - |
1.92 | 9600 | 0.0004 | - | - | - |
1.94 | 9700 | 0.0004 | - | - | - |
1.96 | 9800 | 0.0004 | - | - | - |
1.98 | 9900 | 0.0004 | - | - | - |
2.0 | 10000 | 0.0004 | 0.0004 | 0.8837 | -0.0405 |
2.02 | 10100 | 0.0004 | - | - | - |
2.04 | 10200 | 0.0004 | - | - | - |
2.06 | 10300 | 0.0004 | - | - | - |
2.08 | 10400 | 0.0004 | - | - | - |
2.1 | 10500 | 0.0004 | - | - | - |
2.12 | 10600 | 0.0004 | - | - | - |
2.14 | 10700 | 0.0004 | - | - | - |
2.16 | 10800 | 0.0004 | - | - | - |
2.18 | 10900 | 0.0004 | - | - | - |
2.2 | 11000 | 0.0004 | - | - | - |
2.22 | 11100 | 0.0004 | - | - | - |
2.24 | 11200 | 0.0004 | - | - | - |
2.26 | 11300 | 0.0004 | - | - | - |
2.2800 | 11400 | 0.0004 | - | - | - |
2.3 | 11500 | 0.0004 | - | - | - |
2.32 | 11600 | 0.0004 | - | - | - |
2.34 | 11700 | 0.0004 | - | - | - |
2.36 | 11800 | 0.0004 | - | - | - |
2.38 | 11900 | 0.0004 | - | - | - |
2.4 | 12000 | 0.0004 | - | - | - |
2.42 | 12100 | 0.0004 | - | - | - |
2.44 | 12200 | 0.0004 | - | - | - |
2.46 | 12300 | 0.0004 | - | - | - |
2.48 | 12400 | 0.0004 | - | - | - |
2.5 | 12500 | 0.0004 | - | - | - |
2.52 | 12600 | 0.0004 | - | - | - |
2.54 | 12700 | 0.0004 | - | - | - |
2.56 | 12800 | 0.0004 | - | - | - |
2.58 | 12900 | 0.0004 | - | - | - |
2.6 | 13000 | 0.0004 | - | - | - |
2.62 | 13100 | 0.0004 | - | - | - |
2.64 | 13200 | 0.0004 | - | - | - |
2.66 | 13300 | 0.0004 | - | - | - |
2.68 | 13400 | 0.0004 | - | - | - |
2.7 | 13500 | 0.0004 | - | - | - |
2.7200 | 13600 | 0.0004 | - | - | - |
2.74 | 13700 | 0.0004 | - | - | - |
2.76 | 13800 | 0.0004 | - | - | - |
2.7800 | 13900 | 0.0004 | - | - | - |
2.8 | 14000 | 0.0004 | - | - | - |
2.82 | 14100 | 0.0004 | - | - | - |
2.84 | 14200 | 0.0004 | - | - | - |
2.86 | 14300 | 0.0004 | - | - | - |
2.88 | 14400 | 0.0004 | - | - | - |
2.9 | 14500 | 0.0004 | - | - | - |
2.92 | 14600 | 0.0004 | - | - | - |
2.94 | 14700 | 0.0004 | - | - | - |
2.96 | 14800 | 0.0004 | - | - | - |
2.98 | 14900 | 0.0004 | - | - | - |
3.0 | 15000 | 0.0004 | 0.0004 | 0.8894 | -0.0390 |
Framework Versions
- Python: 3.10.12
- Sentence Transformers: 3.1.1
- Transformers: 4.44.2
- PyTorch: 2.4.1+cu121
- Accelerate: 0.34.2
- Datasets: 2.21.0
- 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",
}
MSELoss
@inproceedings{reimers-2020-multilingual-sentence-bert,
title = "Making Monolingual Sentence Embeddings Multilingual using Knowledge Distillation",
author = "Reimers, Nils and Gurevych, Iryna",
booktitle = "Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing",
month = "11",
year = "2020",
publisher = "Association for Computational Linguistics",
url = "https://arxiv.org/abs/2004.09813",
}
- Downloads last month
- 12
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 alperctnkaya/bge-m3-distilled-en-tr
Base model
BAAI/bge-m3Evaluation results
- Negative Mse on evalself-reported-0.039
- Src2Trg Accuracy on evalself-reported0.895
- Trg2Src Accuracy on evalself-reported0.884
- Mean Accuracy on evalself-reported0.889