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:1614
  - loss:AnglELoss
widget:
  - source_sentence: 姿かたちを変える魔法
    sentences:
      - 昨日夕飯にチキンヌードル食べた？
      - 物の姿を変える魔法が使える村人を知っている？
      - どっちにする？
  - source_sentence: 物の形を変える魔法が使える人
    sentences:
      - どこを探したい？
      - 昨日は何を作ったの？
      - 魔女
  - source_sentence: 花がぬいぐるみに変えられている
    sentences:
      - 誰かが魔法の呪文で花をぬいぐるみに変えた
      - 他にはある？
      - どっちを先にさがしたい？
  - source_sentence: あの木に引っかかってるやつ
    sentences:
      - あれってキミのスカーフ？
      - どんな食べ物が好き？
      - 魔女
  - source_sentence: 井戸を使っていた
    sentences:
      - 夕ご飯は何を食べたの？
      - ぬいぐるみが花
      - 井戸を使った
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: 0.9268897149938042
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.5390470027923584
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.9572153734590284
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.5297622084617615
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.9361702127659575
            name: Cosine Precision
          - type: cosine_recall
            value: 0.9792284866468842
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9451793653848493
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.9256505576208178
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 303.9052734375
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.9564586357039188
            name: Dot F1
          - type: dot_f1_threshold
            value: 300.2573547363281
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.9360795454545454
            name: Dot Precision
          - type: dot_recall
            value: 0.9777448071216617
            name: Dot Recall
          - type: dot_ap
            value: 0.9471967545442288
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.9256505576208178
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 509.136962890625
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.9567099567099567
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 513.729248046875
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.9311797752808989
            name: Manhattan Precision
          - type: manhattan_recall
            value: 0.983679525222552
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9446733439043564
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.9256505576208178
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 23.295352935791016
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.9566473988439306
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 23.295352935791016
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.9323943661971831
            name: Euclidean Precision
          - type: euclidean_recall
            value: 0.9821958456973294
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.9447061843792643
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.9268897149938042
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 509.136962890625
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.9572153734590284
            name: Max F1
          - type: max_f1_threshold
            value: 513.729248046875
            name: Max F1 Threshold
          - type: max_precision
            value: 0.9361702127659575
            name: Max Precision
          - type: max_recall
            value: 0.983679525222552
            name: Max Recall
          - type: max_ap
            value: 0.9471967545442288
            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 = [
    '井戸を使っていた',
    '井戸を使った',
    '夕ご飯は何を食べたの？',
]
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	0.9269
cosine_accuracy_threshold	0.539
cosine_f1	0.9572
cosine_f1_threshold	0.5298
cosine_precision	0.9362
cosine_recall	0.9792
cosine_ap	0.9452
dot_accuracy	0.9257
dot_accuracy_threshold	303.9053
dot_f1	0.9565
dot_f1_threshold	300.2574
dot_precision	0.9361
dot_recall	0.9777
dot_ap	0.9472
manhattan_accuracy	0.9257
manhattan_accuracy_threshold	509.137
manhattan_f1	0.9567
manhattan_f1_threshold	513.7292
manhattan_precision	0.9312
manhattan_recall	0.9837
manhattan_ap	0.9447
euclidean_accuracy	0.9257
euclidean_accuracy_threshold	23.2954
euclidean_f1	0.9566
euclidean_f1_threshold	23.2954
euclidean_precision	0.9324
euclidean_recall	0.9822
euclidean_ap	0.9447
max_accuracy	0.9269
max_accuracy_threshold	509.137
max_f1	0.9572
max_f1_threshold	513.7292
max_precision	0.9362
max_recall	0.9837
max_ap	0.9472

Training Details

Training Dataset

csv

Dataset: csv
Size: 1,614 training samples
Columns: text1, text2, and label
Approximate statistics based on the first 1000 samples:
text1 text2 label
type string string int
details
min: 4 tokens
mean: 8.0 tokens
max: 15 tokens

min: 4 tokens
mean: 8.16 tokens
max: 14 tokens

0: ~18.34%
1: ~81.66%
Samples:

text1 text2 label

見た目を変える魔法 物体を変える 1

昨日何食べたの？ 昨日なに作ったの？ 1

どっち？ どれがおすすめ？ 1

	text1	text2	label
type	string	string	int
details	min: 4 tokens mean: 8.0 tokens max: 15 tokens	min: 4 tokens mean: 8.16 tokens max: 14 tokens	0: ~18.34% 1: ~81.66%

text1	text2	label
`見た目を変える魔法`	`物体を変える`	`1`
`昨日何食べたの？`	`昨日なに作ったの？`	`1`
`どっち？`	`どれがおすすめ？`	`1`

Loss: AnglELoss with these parameters:

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

Evaluation Dataset

csv

Dataset: csv
Size: 1,614 evaluation samples
Columns: text1, text2, and label
Approximate statistics based on the first 1000 samples:
text1 text2 label
type string string int
details
min: 4 tokens
mean: 8.01 tokens
max: 15 tokens

min: 4 tokens
mean: 8.36 tokens
max: 14 tokens

0: ~16.48%
1: ~83.52%
Samples:

text1 text2 label

ぶさいく あほ 1

昨日何食べたの？ 夕ご飯は何を食べたの？ 1

どっちがおすすめ？ どこを探す？ 1

	text1	text2	label
type	string	string	int
details	min: 4 tokens mean: 8.01 tokens max: 15 tokens	min: 4 tokens mean: 8.36 tokens max: 14 tokens	0: ~16.48% 1: ~83.52%

text1	text2	label
`ぶさいく`	`あほ`	`1`
`昨日何食べたの？`	`夕ご飯は何を食べたの？`	`1`
`どっちがおすすめ？`	`どこを探す？`	`1`

Loss: AnglELoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
learning_rate: 2e-05
warmup_ratio: 0.4
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: 3
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.4
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
0	0	-	-	0.9243
1.0	101	1.7051	1.2844	0.9375
2.0	202	0.839	0.7653	0.9472
3.0	303	0.3452	0.7363	0.9472

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",
}

AnglELoss

@misc{li2023angleoptimized,
    title={AnglE-optimized Text Embeddings},
    author={Xianming Li and Jing Li},
    year={2023},
    eprint={2309.12871},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}