metadata

base_model: colorfulscoop/sbert-base-ja
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - cosine_accuracy_threshold
  - cosine_f1
  - cosine_f1_threshold
  - cosine_precision
  - cosine_recall
  - cosine_ap
  - dot_accuracy
  - dot_accuracy_threshold
  - dot_f1
  - dot_f1_threshold
  - dot_precision
  - dot_recall
  - dot_ap
  - manhattan_accuracy
  - manhattan_accuracy_threshold
  - manhattan_f1
  - manhattan_f1_threshold
  - manhattan_precision
  - manhattan_recall
  - manhattan_ap
  - euclidean_accuracy
  - euclidean_accuracy_threshold
  - euclidean_f1
  - euclidean_f1_threshold
  - euclidean_precision
  - euclidean_recall
  - euclidean_ap
  - max_accuracy
  - max_accuracy_threshold
  - max_f1
  - max_f1_threshold
  - max_precision
  - max_recall
  - max_ap
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:680
  - loss:ContrastiveLoss
widget:
  - source_sentence: 両方はだめ？
    sentences:
      - 両方欲しい
      - あほ
      - キッチンを調べよう
  - source_sentence: どっちも欲しくない
    sentences:
      - 誰かが魔法の呪文で花をぬいぐるみに変えた
      - 呪文を試すため
      - 家の中を調べよう
  - source_sentence: この本は？
    sentences:
      - お鍋から匂いがしたから
      - なんでここに本が？
      - 両方行きたい
  - source_sentence: 他のは選べる？
    sentences:
      - 昨日夕飯にチキンヌードル食べた？
      - 別のは選べる？
      - チキンヌードル作った？
  - source_sentence: 猫のぬいぐるみ
    sentences:
      - 両方はだめ？
      - ぬいぐるみ
      - 夜ご飯を食べる前
model-index:
  - name: SentenceTransformer based on colorfulscoop/sbert-base-ja
    results:
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: custom arc semantics data jp
          type: custom-arc-semantics-data-jp
        metrics:
          - type: cosine_accuracy
            value: 0.8897058823529411
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.6581918001174927
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.9044585987261147
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.6180122494697571
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.9466666666666667
            name: Cosine Precision
          - type: cosine_recall
            value: 0.8658536585365854
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9692848872766847
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.8897058823529411
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 374.541748046875
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.9019607843137255
            name: Dot F1
          - type: dot_f1_threshold
            value: 374.541748046875
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.971830985915493
            name: Dot Precision
          - type: dot_recall
            value: 0.8414634146341463
            name: Dot Recall
          - type: dot_ap
            value: 0.9691104975300342
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.8970588235294118
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 453.2839660644531
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.9102564102564101
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 453.2839660644531
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.9594594594594594
            name: Manhattan Precision
          - type: manhattan_recall
            value: 0.8658536585365854
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9687920395428105
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.8897058823529411
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 19.75204086303711
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.9047619047619047
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 23.66771125793457
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.8837209302325582
            name: Euclidean Precision
          - type: euclidean_recall
            value: 0.926829268292683
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.9690811253492324
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.8970588235294118
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 453.2839660644531
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.9102564102564101
            name: Max F1
          - type: max_f1_threshold
            value: 453.2839660644531
            name: Max F1 Threshold
          - type: max_precision
            value: 0.971830985915493
            name: Max Precision
          - type: max_recall
            value: 0.926829268292683
            name: Max Recall
          - type: max_ap
            value: 0.9692848872766847
            name: Max Ap

SentenceTransformer based on colorfulscoop/sbert-base-ja

This is a sentence-transformers model finetuned from colorfulscoop/sbert-base-ja on the csv 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
Base model: colorfulscoop/sbert-base-ja
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 tokens
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': 512, 'do_lower_case': False}) with Transformer model: BertModel 
  (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("sentence_transformers_model_id")
# Run inference
sentences = [
    '猫のぬいぐるみ',
    'ぬいぐるみ',
    '両方はだめ？',
]
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]

Evaluation

Metrics

Binary Classification

Dataset: custom-arc-semantics-data-jp
Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.8897
cosine_accuracy_threshold	0.6582
cosine_f1	0.9045
cosine_f1_threshold	0.618
cosine_precision	0.9467
cosine_recall	0.8659
cosine_ap	0.9693
dot_accuracy	0.8897
dot_accuracy_threshold	374.5417
dot_f1	0.902
dot_f1_threshold	374.5417
dot_precision	0.9718
dot_recall	0.8415
dot_ap	0.9691
manhattan_accuracy	0.8971
manhattan_accuracy_threshold	453.284
manhattan_f1	0.9103
manhattan_f1_threshold	453.284
manhattan_precision	0.9595
manhattan_recall	0.8659
manhattan_ap	0.9688
euclidean_accuracy	0.8897
euclidean_accuracy_threshold	19.752
euclidean_f1	0.9048
euclidean_f1_threshold	23.6677
euclidean_precision	0.8837
euclidean_recall	0.9268
euclidean_ap	0.9691
max_accuracy	0.8971
max_accuracy_threshold	453.284
max_f1	0.9103
max_f1_threshold	453.284
max_precision	0.9718
max_recall	0.9268
max_ap	0.9693

Training Details

Training Dataset

csv

Dataset: csv
Size: 680 training samples
Columns: text1, text2, and label
Approximate statistics based on the first 680 samples:
text1 text2 label
type string string int
details
min: 4 tokens
mean: 8.32 tokens
max: 15 tokens

min: 4 tokens
mean: 8.0 tokens
max: 14 tokens

0: ~41.73%
1: ~58.27%
Samples:

text1 text2 label

試すため ためすため 1

お鍋からの香り お鍋から辛い匂いがしたから 1

なんで話せるの？ なんでしゃべれるの？ 1

	text1	text2	label
type	string	string	int
details	min: 4 tokens mean: 8.32 tokens max: 15 tokens	min: 4 tokens mean: 8.0 tokens max: 14 tokens	0: ~41.73% 1: ~58.27%

text1	text2	label
`試すため`	`ためすため`	`1`
`お鍋からの香り`	`お鍋から辛い匂いがしたから`	`1`
`なんで話せるの？`	`なんでしゃべれるの？`	`1`

Loss: ContrastiveLoss with these parameters:

{
    "distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
    "margin": 0.8,
    "size_average": true
}

Evaluation Dataset

csv

Dataset: csv
Size: 680 evaluation samples
Columns: text1, text2, and label
Approximate statistics based on the first 680 samples:
text1 text2 label
type string string int
details
min: 4 tokens
mean: 8.21 tokens
max: 13 tokens

min: 4 tokens
mean: 8.04 tokens
max: 14 tokens

0: ~39.71%
1: ~60.29%
Samples:

text1 text2 label

村人について教えて 猫のぬいぐるみ 0

ハロー やあ 1

窓から出て行った オブリビオンの魔法 0

	text1	text2	label
type	string	string	int
details	min: 4 tokens mean: 8.21 tokens max: 13 tokens	min: 4 tokens mean: 8.04 tokens max: 14 tokens	0: ~39.71% 1: ~60.29%

text1	text2	label
`村人について教えて`	`猫のぬいぐるみ`	`0`
`ハロー`	`やあ`	`1`
`窓から出て行った`	`オブリビオンの魔法`	`0`

Loss: ContrastiveLoss with these parameters:

{
    "distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
    "margin": 0.8,
    "size_average": true
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
learning_rate: 2e-05
num_train_epochs: 5
warmup_ratio: 0.1
fp16: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 8
per_device_eval_batch_size: 8
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: 2e-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: 0.1
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: 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	loss	custom-arc-semantics-data-jp_max_ap
None	0	-	-	0.9118
1.0	68	0.0481	0.0342	0.9611
2.0	136	0.0307	0.0318	0.9656
3.0	204	0.0218	0.0282	0.9728
4.0	272	0.0169	0.0285	0.9706
5.0	340	0.0144	0.0289	0.9693

Framework Versions

Python: 3.10.14
Sentence Transformers: 3.1.0
Transformers: 4.44.2
PyTorch: 2.4.1+cu121
Accelerate: 0.34.2
Datasets: 2.20.0
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",
}

ContrastiveLoss

@inproceedings{hadsell2006dimensionality,
    author={Hadsell, R. and Chopra, S. and LeCun, Y.},
    booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)},
    title={Dimensionality Reduction by Learning an Invariant Mapping},
    year={2006},
    volume={2},
    number={},
    pages={1735-1742},
    doi={10.1109/CVPR.2006.100}
}