wublewobble/classifier_12

SentenceTransformer based on sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2

This is a sentence-transformers model finetuned from sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2. 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/paraphrase-multilingual-MiniLM-L12-v2
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 384 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': 128, '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})
)

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("wublewobble/classifier_12")
# Run inference
sentences = [
    'Event: Shanghai Old Jazz Band 上海老爵士乐队音乐会\nDescription: Relive the golden era of Shanghai’s jazz scene with this nostalgic concert.\nVenue: Shanghai Music Hall',
    'Concert : Classical Vocals-Asian',
    'Dance : Modern/Contemporary',
]
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: test
• Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.995
cosine_accuracy_threshold	0.5652
cosine_f1	0.7417
cosine_f1_threshold	0.5227
cosine_precision	0.837
cosine_recall	0.6659
cosine_ap	0.7858
cosine_mcc	0.7442

Training Details

Training Dataset

Unnamed Dataset

• Size: 4,140 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 21 tokens mean: 99.72 tokens max: 128 tokens	min: 5 tokens mean: 8.29 tokens max: 21 tokens

• Samples:

  anchor	positive
  Event: The World of Swiss Education and Summer Camps - 2024 [G] Description: Come with your children to discover Switzerland's most esteemed boarding schools, hotel management schools and summer camps on a fun-filled family adventure through the Alps, experience Swiss culture and meet admission directors in person. 3.00pm Doors open. Families are free to discover boarding schools, summer camps & network. Children’s activities begin. 3.10pm Welcome by the Ambassador of Switzerland, HE Frank Grütter. Presentations to introduce Swiss schools and summer camps. 3.30pm Lucky draw 5.00pm Close Venue: The Embassy Room, St. Regis Hotel	Festival/Fair : Business & Professional
  Event: Wine Tasting and Sommelier Experience Description: Join our expert sommelier for an immersive wine tasting experience. Sample premium wines, learn the art of wine pairing, and develop a deeper appreciation for fine wines in an intimate setting. Venue: Wine Tasting Room	Lifestyle/Leisure : Service
  Event: Huayi 华艺节 2020 Storytellers' Wisdom - A Crosstalk Production 十五万大军直取西城而来 Description: What does Detective Conan and Justice Bao have in common? Can the capable Sun Wukong with his endless transformations survive in the modern society? What can Jin Yong’s stories tell you about the philosophy of ‘three’? With a focus on the Empty Fort Strategy from the classic Chinese military directives Thirty-Six Stratagems , Storytellers’ Wisdom is a lighthearted crosstalk production that enacts the various chapters of Chinese culture and history through an engaging performance filled with clever dialogue and witty humour. An original creation by the renowned Comedians Workshop from Taiwan, Storytellers’ Wisdom features a selection of the group’s best works performed by established theatre practitioners including Feng Yi-Gang and Sung Shao-Ching. Discover humorous anecdotes about life told through stories from Romance of the Three Kingdoms , Justice Bao, Sun Wukong and classics fro...	Theatre : Comedy

• Loss: CachedMultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 92
• per_device_eval_batch_size: 92
• num_train_epochs: 10
• warmup_ratio: 0.1
• 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: 92
• per_device_eval_batch_size: 92
• 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: 10
• 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: 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	test_cosine_ap
0	0	-	0.1384
1.1111	50	2.0041	0.5899
2.2222	100	1.0715	0.6977
3.3333	150	0.668	0.7221
4.4444	200	0.4198	0.7442
5.5556	250	0.2544	0.7490
6.6667	300	0.1533	0.7736
7.7778	350	0.0994	0.7806
8.8889	400	0.066	0.7834
10.0	450	0.0491	0.7858

Framework Versions

• Python: 3.11.11
• Sentence Transformers: 3.4.1
• Transformers: 4.48.3
• PyTorch: 2.6.0+cu124
• Accelerate: 1.3.0
• Datasets: 3.4.1
• Tokenizers: 0.21.1

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",
}

CachedMultipleNegativesRankingLoss

@misc{gao2021scaling,
    title={Scaling Deep Contrastive Learning Batch Size under Memory Limited Setup},
    author={Luyu Gao and Yunyi Zhang and Jiawei Han and Jamie Callan},
    year={2021},
    eprint={2101.06983},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}