metadata

base_model: sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
language:
  - hu
library_name: sentence-transformers
license: apache-2.0
metrics:
  - cosine_accuracy
  - dot_accuracy
  - manhattan_accuracy
  - euclidean_accuracy
  - max_accuracy
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:1044013
  - loss:MultipleNegativesRankingLoss
widget:
  - source_sentence: Emberek várnak a lámpánál kerékpárral.
    sentences:
      - Az emberek piros lámpánál haladnak.
      - Az emberek a kerékpárjukon vannak.
      - Egy fekete kutya úszik a vízben egy teniszlabdával a szájában
  - source_sentence: A kutya a vízben van.
    sentences:
      - >-
        Két férfi takarítja a havat a tetőről, az egyik egy emelőben ül, a másik
        pedig a tetőn.
      - A macska a vízben van, és dühös.
      - Egy kutya van a vízben, a szájában egy faág.
  - source_sentence: A nő feketét visel.
    sentences:
      - Egy barna kutya fröcsköl, ahogy úszik a vízben.
      - Egy tetoválással rendelkező nő, aki fekete tank tetején néz a földre.
      - >-
        Egy kékbe öltözött nő intenzív arckifejezéssel üti a teniszlabdát. A
        képen:
  - source_sentence: Az emberek alszanak.
    sentences:
      - Három ember beszélget egy városi utcán.
      - A nő fehéret visel.
      - Egy apa és a fia ölelgeti alvás közben.
  - source_sentence: Az emberek alszanak.
    sentences:
      - >-
        Egy feketébe öltözött nő cigarettát és bevásárlótáskát tart a kezében,
        miközben egy idősebb nő átmegy az utcán.
      - >-
        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.
      - Egy apa és a fia ölelgeti alvás közben.
model-index:
  - name: paraphrase-multilingual-MiniLM-L12-v2-hu
    results:
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: all nli dev
          type: all-nli-dev
        metrics:
          - type: cosine_accuracy
            value: 0.9918
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.0102
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 0.99
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.99
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 0.9918
            name: Max Accuracy
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: all nli test
          type: all-nli-test
        metrics:
          - type: cosine_accuracy
            value: 0.9937878787878788
            name: Cosine Accuracy
          - type: dot_accuracy
            value: 0.00803030303030303
            name: Dot Accuracy
          - type: manhattan_accuracy
            value: 0.9928787878787879
            name: Manhattan Accuracy
          - type: euclidean_accuracy
            value: 0.9924242424242424
            name: Euclidean Accuracy
          - type: max_accuracy
            value: 0.9937878787878788
            name: Max Accuracy

paraphrase-multilingual-MiniLM-L12-v2-hu

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-v2")
# 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	0.9918
dot_accuracy	0.0102
manhattan_accuracy	0.99
euclidean_accuracy	0.99
max_accuracy	0.9918

Triplet

Dataset: all-nli-test
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.9938
dot_accuracy	0.008
manhattan_accuracy	0.9929
euclidean_accuracy	0.9924
max_accuracy	0.9938

Training Details

Training Dataset

train

Dataset: train
Size: 1,044,013 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	-
0.0123	100	2.5472	-	-	-
0.0245	200	2.0478	-	-	-
0.0368	300	1.8426	-	-	-
0.0490	400	1.698	-	-	-
0.0613	500	1.5715	-	-	-
0.0736	600	1.4616	-	-	-
0.0858	700	1.6106	-	-	-
0.0981	800	1.5849	-	-	-
0.1103	900	1.5374	-	-	-
0.1226	1000	1.3653	-	-	-
0.1349	1100	1.274	-	-	-
0.1471	1200	1.1907	-	-	-
0.1594	1300	1.2155	-	-	-
0.1716	1400	1.2786	-	-	-
0.1839	1500	1.1062	-	-	-
0.1962	1600	1.0289	-	-	-
0.2084	1700	1.0013	-	-	-
0.2207	1800	0.9209	-	-	-
0.2329	1900	0.8095	-	-	-
0.2452	2000	0.9753	0.1916	0.9558	-
0.2574	2100	0.8728	-	-	-
0.2697	2200	0.8343	-	-	-
0.2820	2300	0.7203	-	-	-
0.2942	2400	0.6901	-	-	-
0.3065	2500	0.6606	-	-	-
0.3187	2600	0.7205	-	-	-
0.3310	2700	0.7479	-	-	-
0.3433	2800	0.6677	-	-	-
0.3555	2900	1.2531	-	-	-
0.3678	3000	1.3619	-	-	-
0.3800	3100	1.3923	-	-	-
0.3923	3200	1.412	-	-	-
0.4046	3300	1.3904	-	-	-
0.4168	3400	1.3782	-	-	-
0.4291	3500	1.3601	-	-	-
0.4413	3600	1.3582	-	-	-
0.4536	3700	1.3402	-	-	-
0.4659	3800	1.32	-	-	-
0.4781	3900	1.3277	-	-	-
0.4904	4000	1.3112	0.0699	0.987	-
0.5026	4100	1.2992	-	-	-
0.5149	4200	1.3005	-	-	-
0.5272	4300	1.2978	-	-	-
0.5394	4400	1.272	-	-	-
0.5517	4500	1.2864	-	-	-
0.5639	4600	1.2519	-	-	-
0.5762	4700	1.1924	-	-	-
0.5885	4800	1.1778	-	-	-
0.6007	4900	1.1801	-	-	-
0.6130	5000	1.1666	-	-	-
0.6252	5100	1.1682	-	-	-
0.6375	5200	1.1518	-	-	-
0.6497	5300	1.1606	-	-	-
0.6620	5400	1.1534	-	-	-
0.6743	5500	1.1473	-	-	-
0.6865	5600	1.1596	-	-	-
0.6988	5700	1.1536	-	-	-
0.7110	5800	1.1517	-	-	-
0.7233	5900	1.1517	-	-	-
0.7356	6000	1.153	0.0359	0.9896	-
0.7478	6100	1.142	-	-	-
0.7601	6200	1.093	-	-	-
0.7723	6300	1.1764	-	-	-
0.7846	6400	1.1868	-	-	-
0.7969	6500	1.0308	-	-	-
0.8091	6600	1.0122	-	-	-
0.8214	6700	1.0084	-	-	-
0.8336	6800	1.0151	-	-	-
0.8459	6900	1.0121	-	-	-
0.8582	7000	1.0071	-	-	-
0.8704	7100	1.1543	-	-	-
0.8827	7200	1.1915	-	-	-
0.8949	7300	1.2224	-	-	-
0.9072	7400	1.1463	-	-	-
0.9195	7500	1.0254	-	-	-
0.9317	7600	1.2396	-	-	-
0.9440	7700	1.1225	-	-	-
0.9562	7800	0.7177	-	-	-
0.9685	7900	0.0681	-	-	-
0.9808	8000	0.0264	0.0317	0.9918	-
0.9930	8100	0.078	-	-	-
1.0	8157	-	-	-	0.9938

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: 3.0.2
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}
}