x2bee/ModernBERT-SimCSE-multitask_v04

SentenceTransformer based on CocoRoF/ModernBERT-SimCSE_v04

This is a sentence-transformers model finetuned from CocoRoF/ModernBERT-SimCSE_v04. 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: CocoRoF/ModernBERT-SimCSE_v04
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 dimensions
• Similarity Function: Cosine Similarity

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: 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})
  (2): Dense({'in_features': 768, 'out_features': 768, 'bias': True, 'activation_function': 'torch.nn.modules.activation.Tanh'})
)

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("CocoRoF/ModernBERT-SimCSE-multitask_v04")
# Run inference
sentences = [
    '버스가 바쁜 길을 따라 운전한다.',
    '녹색 버스가 도로를 따라 내려간다.',
    '그 여자는 데이트하러 가는 중이다.',
]
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

Semantic Similarity

• Dataset: sts_dev
• Evaluated with EmbeddingSimilarityEvaluator

Metric	Value
pearson_cosine	0.7847
spearman_cosine	0.7871
pearson_euclidean	0.7259
spearman_euclidean	0.7209
pearson_manhattan	0.7252
spearman_manhattan	0.7203
pearson_dot	0.621
spearman_dot	0.6255
pearson_max	0.7847
spearman_max	0.7871

Training Details

Training Dataset

Unnamed Dataset

• Size: 5,749 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 7 tokens mean: 12.69 tokens max: 31 tokens	min: 7 tokens mean: 12.56 tokens max: 27 tokens	min: 0.0 mean: 0.45 max: 1.0

• Samples:

  sentence1	sentence2	score
  비행기가 이륙하고 있다.	비행기가 이륙하고 있다.	1.0
  한 남자가 큰 플루트를 연주하고 있다.	남자가 플루트를 연주하고 있다.	0.76
  한 남자가 피자에 치즈를 뿌려놓고 있다.	한 남자가 구운 피자에 치즈 조각을 뿌려놓고 있다.	0.76

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Evaluation Dataset

Unnamed Dataset

• Size: 1,500 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	min: 6 tokens mean: 18.89 tokens max: 51 tokens	min: 5 tokens mean: 18.92 tokens max: 50 tokens	min: 0.0 mean: 0.42 max: 1.0

• Samples:

  sentence1	sentence2	score
  안전모를 가진 한 남자가 춤을 추고 있다.	안전모를 쓴 한 남자가 춤을 추고 있다.	1.0
  어린아이가 말을 타고 있다.	아이가 말을 타고 있다.	0.95
  한 남자가 뱀에게 쥐를 먹이고 있다.	남자가 뱀에게 쥐를 먹이고 있다.	1.0

• Loss: CosineSimilarityLoss with these parameters:

  {
      "loss_fct": "torch.nn.modules.loss.MSELoss"
  }
  

Training Hyperparameters

Non-Default Hyperparameters

• overwrite_output_dir: True
• eval_strategy: steps
• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• gradient_accumulation_steps: 8
• learning_rate: 1e-05
• num_train_epochs: 10.0
• warmup_ratio: 0.1
• push_to_hub: True
• hub_model_id: CocoRoF/ModernBERT-SimCSE-multitask_v04
• hub_strategy: checkpoint
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: True
• do_predict: False
• eval_strategy: steps
• 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: 8
• eval_accumulation_steps: None
• torch_empty_cache_steps: None
• learning_rate: 1e-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: 10.0
• 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: 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: True
• 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: True
• resume_from_checkpoint: None
• hub_model_id: CocoRoF/ModernBERT-SimCSE-multitask_v04
• hub_strategy: checkpoint
• 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	sts_dev_spearman_max
0.2228	10	0.0285	-	-
0.4457	20	0.0396	-	-
0.6685	30	0.0396	0.0376	0.7647
0.8914	40	0.0594	-	-
1.1337	50	0.0438	-	-
1.3565	60	0.0302	0.0358	0.7723
1.5794	70	0.0398	-	-
1.8022	80	0.0457	-	-
2.0446	90	0.0464	0.0347	0.7805
2.2674	100	0.026	-	-
2.4903	110	0.0331	-	-
2.7131	120	0.0318	0.0329	0.7837
2.9359	130	0.0399	-	-
3.1783	140	0.0264	-	-
3.4011	150	0.0268	0.0332	0.7884
3.6240	160	0.0241	-	-
3.8468	170	0.0309	-	-
4.0891	180	0.0263	0.0326	0.7918
4.3120	190	0.0164	-	-
4.5348	200	0.0226	-	-
4.7577	210	0.0196	0.0314	0.7896
4.9805	220	0.0217	-	-
5.2228	230	0.0134	-	-
5.4457	240	0.0157	0.0320	0.7911
5.6685	250	0.0136	-	-
5.8914	260	0.0143	-	-
6.1337	270	0.0114	0.0322	0.7907
6.3565	280	0.0077	-	-
6.5794	290	0.0116	-	-
6.8022	300	0.0087	0.0313	0.7868
7.0446	310	0.0088	-	-
7.2674	320	0.0048	-	-
7.4903	330	0.0068	0.0317	0.7895
7.7131	340	0.006	-	-
7.9359	350	0.0051	-	-
8.1783	360	0.0039	0.0323	0.7882
8.4011	370	0.0036	-	-
8.6240	380	0.0045	-	-
8.8468	390	0.0032	0.0317	0.7841
9.0891	400	0.0031	-	-
9.3120	410	0.0021	-	-
9.5348	420	0.0029	0.0323	0.7871
9.7577	430	0.0023	-	-
9.9805	440	0.0027	-	-

Framework Versions

• Python: 3.11.10
• Sentence Transformers: 3.3.1
• Transformers: 4.48.0.dev0
• PyTorch: 2.5.1+cu121
• Accelerate: 1.1.0
• Datasets: 3.1.0
• 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",
}