karsar/paraphrase-multilingual-MiniLM-L12-hu

e5-base_hun

This is a sentence-transformers model finetuned from sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 on the train 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 tokens
• Similarity Function: Cosine Similarity
• Training Dataset:
  • train
• Language: hu
• License: apache-2.0

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("karsar/paraphrase-multilingual-MiniLM-L12-hu")
# Run inference
sentences = [
    'Az emberek alszanak.',
    'Egy apa és a fia ölelgeti alvás közben.',
    'Egy csoport ember ül egy nyitott, térszerű területen, mögötte nagy bokrok és egy sor viktoriánus stílusú épület, melyek közül sokat a kép jobb oldalán lévő erős elmosódás tesz kivehetetlenné.',
]
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

Triplet

• Dataset: all-nli-dev
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9746
dot_accuracy	0.0284
manhattan_accuracy	0.9676
euclidean_accuracy	0.9658
max_accuracy	0.9746

Triplet

• Dataset: all-nli-test
• Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9921
dot_accuracy	0.0086
manhattan_accuracy	0.9897
euclidean_accuracy	0.9895
max_accuracy	0.9921

Training Details

Training Dataset

train

• Dataset: train
• Size: 457,856 training samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 7 tokens mean: 11.73 tokens max: 56 tokens	min: 6 tokens mean: 15.24 tokens max: 47 tokens	min: 7 tokens mean: 16.07 tokens max: 53 tokens

• Samples:

  anchor	positive	negative
  Egy lóháton ülő ember átugrik egy lerombolt repülőgép felett.	Egy ember a szabadban, lóháton.	Egy ember egy étteremben van, és omlettet rendel.
  Gyerekek mosolyogva és integetett a kamera	Gyermekek vannak jelen	A gyerekek homlokot rántanak
  Egy fiú ugrál a gördeszkát a közepén egy piros híd.	A fiú gördeszkás trükköt csinál.	A fiú korcsolyázik a járdán.

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Evaluation Dataset

train

• Dataset: train
• Size: 5,000 evaluation samples
• Columns: anchor, positive, and negative
• Approximate statistics based on the first 1000 samples:

  	anchor	positive	negative
  type	string	string	string
  details	min: 7 tokens mean: 11.73 tokens max: 56 tokens	min: 6 tokens mean: 15.24 tokens max: 47 tokens	min: 7 tokens mean: 16.07 tokens max: 53 tokens

• Samples:

  anchor	positive	negative
  Egy lóháton ülő ember átugrik egy lerombolt repülőgép felett.	Egy ember a szabadban, lóháton.	Egy ember egy étteremben van, és omlettet rendel.
  Gyerekek mosolyogva és integetett a kamera	Gyermekek vannak jelen	A gyerekek homlokot rántanak
  Egy fiú ugrál a gördeszkát a közepén egy piros híd.	A fiú gördeszkás trükköt csinál.	A fiú korcsolyázik a járdán.

• Loss: MultipleNegativesRankingLoss with these parameters:

  {
      "scale": 20.0,
      "similarity_fct": "cos_sim"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 128
• per_device_eval_batch_size: 128
• num_train_epochs: 1
• warmup_ratio: 0.1
• bf16: 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: 128
• per_device_eval_batch_size: 128
• 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: 1
• 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: True
• fp16: False
• 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: False
• hub_always_push: False
• gradient_checkpointing: False
• gradient_checkpointing_kwargs: None
• include_inputs_for_metrics: False
• 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
• eval_use_gather_object: False
• batch_sampler: no_duplicates
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	train loss	all-nli-dev_max_accuracy	all-nli-test_max_accuracy
0	0	-	-	0.7574	-
0.0280	100	2.3495	-	-	-
0.0559	200	1.8588	-	-	-
0.0839	300	1.7156	-	-	-
0.1118	400	1.609	-	-	-
0.1398	500	1.5286	-	-	-
0.1677	600	1.4425	-	-	-
0.1957	700	1.6016	-	-	-
0.2237	800	1.5278	-	-	-
0.2516	900	1.4255	-	-	-
0.2796	1000	1.2084	-	-	-
0.3075	1100	1.1248	-	-	-
0.3355	1200	1.0773	-	-	-
0.3634	1300	1.1373	-	-	-
0.3914	1400	1.222	-	-	-
0.4193	1500	1.048	-	-	-
0.4473	1600	0.9319	-	-	-
0.4753	1700	0.8837	-	-	-
0.5032	1800	0.8402	-	-	-
0.5312	1900	0.7515	-	-	-
0.5591	2000	0.9405	0.1310	0.9746	-
0.5871	2100	0.8526	-	-	-
0.6150	2200	0.7886	-	-	-
0.6430	2300	0.6704	-	-	-
0.6710	2400	0.6488	-	-	-
0.6989	2500	0.635	-	-	-
0.7269	2600	0.7242	-	-	-
0.7548	2700	0.7593	-	-	-
0.7828	2800	0.62	-	-	-
0.8107	2900	0.4302	-	-	-
0.8387	3000	0.2952	-	-	-
0.8666	3100	0.3354	-	-	-
0.8946	3200	0.3221	-	-	-
0.9226	3300	0.4317	-	-	-
0.9505	3400	0.3185	-	-	-
0.9785	3500	0.433	-	-	-
1.0	3577	-	-	-	0.9921

Framework Versions

• Python: 3.11.8
• Sentence Transformers: 3.1.1
• Transformers: 4.44.0
• PyTorch: 2.3.0.post101
• Accelerate: 0.33.0
• Datasets: 2.18.0
• Tokenizers: 0.19.0

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",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}