metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:478146
  - loss:CoSENTLoss
widget:
  - source_sentence: >-
      However, its underutilization is mainly due to the absence of a concrete
      and coherent dissemination strategy.
    sentences:
      - >-
        At the same time, they need to understand that living in Europe brings
        great responsibilities in addition to great benefits.
      - >-
        The mainstay of any intelligent and patriotic mineral policy can be
        summed up in the following postulate: "since minerals are exhaustible,
        they should only be exploited with the maximum return for the economy of
        the country where they are mined".
      - >-
        We must move quickly to a shared sustainable energy supply, sustainable
        transportation and clean air.
  - source_sentence: >-
      Their track record shows they do not support Australia<92>s traditional
      industries because they are constantly pandering to the Greens.
    sentences:
      - >-
        An economic dynamic based on the sustainable development of national
        potential, equitable access to the means of production, social justice,
        environmental conservation, the incorporation of added value, the
        promotion of competitiveness and self-management,
      - >-
        the cry "El campo no aguanta más" (The countryside can't take it
        anymore), of the peasant movement and its proclamation of "Salvemos al
        Campo para salvar a México" (Let's save the countryside to save Mexico);
      - >-
        On the other hand, increasing defence capacity is directly related to
        the involvement of all citizens in appropriate programmes, which,
        together with the acquisition of skills, experience and organisation,
        also contribute to forging a spirit of militancy and collectivity.
  - source_sentence: >-
      We will prepare the proposals of the United Nations Declaration on the
      Rights of the Child in line with the commitments made.
    sentences:
      - >-
        For the presentation of Czech culture, we will also use the upcoming
        major anniversaries (100 years of the founding of Czechoslovakia, the
        30th anniversary of the canonization of Agnes of Bohemia, 600 years
        since the birth of George of Poděbrady, etc.).
      - >-
        Separate prison units for young people should be established, and
        special rehabilitation measures should be introduced in these units.
      - >-
        Austrian citizenship is a valuable asset and should not become
        accessible to those who do not abide by the laws of our state.
  - source_sentence: >-
      Third, CD&V wants to strengthen the social sustainability of our
      agriculture and horticulture sector.
    sentences:
      - >-
        We will take a farm-level approach where possible so that low-emissions
        farmers are rewarded with a lower cost through the ETS, rather than the
        current approach that assumes each cow, for instance, has the same
        emissions on every farm.
      - >-
        In addition, 20 billion euros in tax revenues are fraudulently evaded
        every year (the equivalent of the healthcare budget).
      - >-
        87 percent of arrested undocumented migrants are released sooner or
        later, but without papers, in a lawless situation.
  - source_sentence: >-
      This incites social hatred, threatens economic and social stability, and
      undermines trust in the authorities.
    sentences:
      - "\_The conditions for a healthy entrepreneurship, where the most innovative and creative win and where the source of enrichment cannot be property speculation or guilds and networks.   "
      - >-
        According to statistics from the Attorney General's Office, since
        February 2005, when the implementation of the PSD was announced, the
        rate of violent deaths per 100,000 inhabitants has dropped from 26.41 in
        December 2005 to 18.43 in December 2007.
      - >-
        As a result, the profits of the oligarchs are more than 400 times what
        our entire country gets from the exploitation of natural resources.
pipeline_tag: sentence-similarity
library_name: sentence-transformers

SentenceTransformer

This is a sentence-transformers model trained on the json dataset. It maps sentences & paragraphs to a 768-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
Maximum Sequence Length: 8192 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- json

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': 8192, 'do_lower_case': False}) with Transformer model: ModernBertModel 
  (1): Pooling({'word_embedding_dimension': 768, '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("LequeuISIR/final-DPR-8e-05")
# Run inference
sentences = [
    'This incites social hatred, threatens economic and social stability, and undermines trust in the authorities.',
    '\xa0The conditions for a healthy entrepreneurship, where the most innovative and creative win and where the source of enrichment cannot be property speculation or guilds and networks.   ',
    'As a result, the profits of the oligarchs are more than 400 times what our entire country gets from the exploitation of natural resources.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Training Details

Training Dataset

json

Dataset: json
Size: 478,146 training samples
Columns: sentence1, sentence2, and label
Approximate statistics based on the first 1000 samples:
sentence1 sentence2 label
type string string int
details
min: 17 tokens
mean: 33.73 tokens
max: 107 tokens

min: 16 tokens
mean: 33.84 tokens
max: 101 tokens

0: ~57.50%
1: ~4.10%
2: ~38.40%

	sentence1	sentence2	label
type	string	string	int
details	min: 17 tokens mean: 33.73 tokens max: 107 tokens	min: 16 tokens mean: 33.84 tokens max: 101 tokens	0: ~57.50% 1: ~4.10% 2: ~38.40%

Samples:

sentence1	sentence2	label
`There have also been other important structural changes in the countryside, which have come together to form this new, as yet unknown, country.`	`Meanwhile, investment, which is the way to increase production, employment capacity and competitiveness of the economy, fell from 20% of output in 1974 to only 11.8% on average between 1984 and 1988.`	`0`
`Introduce new visa categories so we can be responsive to humanitarian needs and incentivise greater investment in our domestic infrastructure and regional economies`	`The purpose of the project is to design and implement public policies aimed at achieving greater and faster inclusion of immigrants.`	`2`
`and economic crimes that seriously and generally affect the fundamental rights of individuals and the international community as a whole.`	`For the first time in the history, not only of Ecuador, but of the entire world, a government promoted a public audit process of the foreign debt and declared some of its tranches illegitimate and immoral.`	`0`

Loss: CoSENTLoss with these parameters:

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

Evaluation Dataset

json

Dataset: json
Size: 478,146 evaluation samples
Columns: sentence1, sentence2, and label
Approximate statistics based on the first 1000 samples:
sentence1 sentence2 label
type string string int
details
min: 17 tokens
mean: 33.62 tokens
max: 103 tokens

min: 16 tokens
mean: 34.48 tokens
max: 111 tokens

0: ~57.30%
1: ~2.90%
2: ~39.80%

	sentence1	sentence2	label
type	string	string	int
details	min: 17 tokens mean: 33.62 tokens max: 103 tokens	min: 16 tokens mean: 34.48 tokens max: 111 tokens	0: ~57.30% 1: ~2.90% 2: ~39.80%

Samples:

sentence1	sentence2	label
`The anchoring of the Slovak Republic in the European Union allows citizens to feel: secure politically, secure economically, secure socially.`	`Radikale Venstre wants Denmark to participate fully and firmly in EU cooperation on immigration, asylum and cross-border crime.`	`2`
`Portugal's participation in the Community's negotiation of the next financial perspective should also be geared in the same direction.`	`Given the dynamic international framework, safeguarding the national interest requires adjustments to each of these vectors.`	`2`
`On asylum, the Green Party will: Dismantle the direct provision system and replace it with an efficient and humane system for determining the status of asylum seekers`	`The crisis in the coal sector subsequently forced these immigrant workers to move into other economic sectors such as metallurgy, chemicals, construction and transport.`	`2`

Loss: CoSENTLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 64
per_device_eval_batch_size: 64
learning_rate: 8e-05
num_train_epochs: 5
warmup_ratio: 0.05
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: 64
per_device_eval_batch_size: 64
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: 8e-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: 5
max_steps: -1
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
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: None
hub_always_push: False
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
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
use_liger_kernel: False
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.0837	500	0.7889	9.5828
0.1673	1000	1.2158	9.3274
0.2510	1500	1.8215	9.4274
0.3346	2000	2.3548	8.2583
0.4183	2500	2.7493	8.1446
0.5019	3000	2.8998	7.9046
0.5856	3500	2.9298	8.0640
0.6692	4000	2.9053	7.2746
0.7529	4500	3.0905	7.5099
0.8365	5000	3.1864	7.3883
0.9202	5500	3.2322	6.9968
1.0038	6000	3.1194	7.4682
1.0875	6500	3.0122	7.7295
1.1712	7000	3.0453	7.1696
1.2548	7500	2.9439	7.2775
1.3385	8000	3.1108	7.4838
1.4221	8500	2.8512	7.5204
1.5058	9000	2.9865	7.4528
1.5894	9500	2.9995	8.0682
1.6731	10000	3.1073	7.5344
1.7567	10500	3.0631	7.4572
1.8404	11000	2.9915	7.4961
1.9240	11500	3.0445	7.3575
2.0077	12000	2.9501	7.9786
2.0914	12500	2.3377	8.6208
2.1750	13000	2.2833	8.8356
2.2587	13500	2.2785	8.8709
2.3423	14000	2.3012	8.6250
2.4260	14500	2.3488	8.1099
2.5096	15000	2.095	9.2305
2.5933	15500	2.4123	8.6405
2.6769	16000	2.2236	8.7805
2.7606	16500	2.3367	8.7110
2.8442	17000	2.1159	8.6447
2.9279	17500	2.1622	8.7123
3.0115	18000	2.1916	9.0314
3.0952	18500	1.604	9.3373
3.1789	19000	1.4116	9.6509
3.2625	19500	1.4036	9.9127
3.3462	20000	1.5392	9.8093
3.4298	20500	1.5791	9.8325
3.5135	21000	1.5343	9.7822
3.5971	21500	1.3913	9.6243
3.6808	22000	1.5151	9.9644
3.7644	22500	1.3922	9.7816
3.8481	23000	1.3361	9.5338
3.9317	23500	1.3363	9.8282
4.0154	24000	1.2234	10.2117
4.0990	24500	0.5927	10.4107
4.1827	25000	0.6879	10.4405
4.2664	25500	0.6832	10.5138
4.3500	26000	0.6514	10.2798
4.4337	26500	0.7396	10.3250
4.5173	27000	0.6813	10.4115
4.6010	27500	0.765	10.1365
4.6846	28000	0.5915	10.2402
4.7683	28500	0.5028	10.3197
4.8519	29000	0.5306	10.3270
4.9356	29500	0.5886	10.3543

Framework Versions

Python: 3.9.21
Sentence Transformers: 3.4.0
Transformers: 4.48.1
PyTorch: 2.5.1+cu124
Accelerate: 1.3.0
Datasets: 3.2.0
Tokenizers: 0.21.0

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",
}

CoSENTLoss

@online{kexuefm-8847,
    title={CoSENT: A more efficient sentence vector scheme than Sentence-BERT},
    author={Su Jianlin},
    year={2022},
    month={Jan},
    url={https://kexue.fm/archives/8847},
}