metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:19598
  - loss:CoSENTLoss
base_model: sentence-transformers/all-MiniLM-L6-v2
widget:
  - source_sentence: soi tỷ lệ Southampton vs Nottingham (21h00, 24/8), vòng 2 Ngoại hạng Anh
    sentences:
      - nhận định Oakleigh Cannons vs Macarthur
      - dự đoán Mallorca vs Athletic Bilbao
      - soi kèo Southampton vs Nottingham Forest
  - source_sentence: Melbourne Victory vs Macarthur 12h00 ngày 3/11 (VĐQG Australia 2024/25).
    sentences:
      - tỷ lệ Tijuana vs Leon
      - 'Melbourne Victory vs Brisbane Roar  '
      - Hải Phòng vs SHB Đà Nẵng
  - source_sentence: Banfield vs Estudiantes 4h00 ngày 8/10 (VĐQG Argentina 2024).
    sentences:
      - arsenal vs psg
      - Shandong Luneng vs Qingdaoyangcheng
      - 'Boca Juniors vs River Plate  '
  - source_sentence: 'St Pauli vs Bayern Munich (21h30 ngày 9/11): Khó có bất ngờ.'
    sentences:
      - st pauli vs bayern munich
      - Seattle Sounders vs Houston Dynamo
      - kyrgyzstan vs triều tiên
  - source_sentence: 'Juventus vs Napoli (23h00 ngày 21/9): Không dễ cho chủ nhà.'
    sentences:
      - cruz azul vs juarez
      - Real Madrid vs Barcelona
      - El Salvador vs Montserrat
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - cosine_accuracy_threshold
  - cosine_f1
  - cosine_f1_threshold
  - cosine_precision
  - cosine_recall
  - cosine_ap
model-index:
  - name: SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
    results:
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: sport query title dev
          type: sport_query_title_dev
        metrics:
          - type: cosine_accuracy
            value: 0.9943877551020408
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.6410836577415466
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.9943269726663229
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.6107593178749084
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.9958677685950413
            name: Cosine Precision
          - type: cosine_recall
            value: 0.9927909371781668
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9995956398472251
            name: Cosine Ap

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2 on the csv 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/all-MiniLM-L6-v2
Maximum Sequence Length: 256 tokens
Output Dimensionality: 384 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- csv

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': 256, '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})
  (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("Tam1032/MiniLM6-v2-sport")
# Run inference
sentences = [
    'Juventus vs Napoli (23h00 ngày 21/9): Không dễ cho chủ nhà.',
    'Real Madrid vs Barcelona',
    'El Salvador vs Montserrat',
]
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

Binary Classification

Dataset: sport_query_title_dev
Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.9944
cosine_accuracy_threshold	0.6411
cosine_f1	0.9943
cosine_f1_threshold	0.6108
cosine_precision	0.9959
cosine_recall	0.9928
cosine_ap	0.9996

Training Details

Training Dataset

csv

Dataset: csv
Size: 19,598 training samples
Columns: hypothesis, premise, and label
Approximate statistics based on the first 1000 samples:
hypothesis premise label
type string string int
details
min: 12 tokens
mean: 27.44 tokens
max: 37 tokens

min: 5 tokens
mean: 9.63 tokens
max: 55 tokens

0: ~50.20%
1: ~49.80%

	hypothesis	premise	label
type	string	string	int
details	min: 12 tokens mean: 27.44 tokens max: 37 tokens	min: 5 tokens mean: 9.63 tokens max: 55 tokens	0: ~50.20% 1: ~49.80%

Samples:

hypothesis	premise	label
`bóng đá Las Palmas vs Girona, 23h30 ngày 26/10: Trừng phạt chủ nhà.`	`Las Palmas vs Girona`	`1`
`Seattle Sounders vs Houston Dynamo 9h30 ngày 29/9 (Nhà nghề Mỹ 2024).`	`dự đoán Seattle Sounders vs Houston Dynamo`	`1`
`bóng đá Tây Ban Nha vs Đan Mạch, 01h45 ngày 13/10: Khuất phục ‘lính chì’.`	`bóng đá Tây Ban Nha vs Đan Mạch`	`1`

Loss: CoSENTLoss with these parameters:

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

Evaluation Dataset

csv

Dataset: csv
Size: 19,598 evaluation samples
Columns: hypothesis, premise, and label
Approximate statistics based on the first 1000 samples:
hypothesis premise label
type string string int
details
min: 12 tokens
mean: 27.15 tokens
max: 40 tokens

min: 4 tokens
mean: 9.55 tokens
max: 40 tokens

0: ~51.40%
1: ~48.60%

	hypothesis	premise	label
type	string	string	int
details	min: 12 tokens mean: 27.15 tokens max: 40 tokens	min: 4 tokens mean: 9.55 tokens max: 40 tokens	0: ~51.40% 1: ~48.60%

Samples:

hypothesis	premise	label
`Hải Phòng vs CAHN (19h15 ngày 15/9): Điểm tựa sân nhà.`	`kết quả Hải Phòng vs CAHN`	`1`
`Kuwait vs Jordan 1h15 ngày 20/11 (Vòng loại World Cup 2026).`	`Kuwait vs Iraq`	`0`
`bóng đá Parma vs Empoli 18h30 ngày 27/10 (Serie A 2024/25).`	`nhận định Parma vs Empoli`	`1`

Loss: CoSENTLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
warmup_ratio: 0.1
fp16: True

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
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: 3
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: False
fp16: True
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: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss	sport_query_title_dev_cosine_ap
1.0	1103	-	0.1376	0.9991
1.4506	1600	0.3994	-	-
2.0	2206	-	0.0693	0.9994
2.9012	3200	0.0442	-	-
3.0	3309	-	0.0534	0.9996

Framework Versions

Python: 3.11.7
Sentence Transformers: 3.3.1
Transformers: 4.47.0
PyTorch: 2.5.1+cu124
Accelerate: 1.2.1
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},
}