RomainDarous/fine_tuned_generalized

SentenceTransformer based on RomainDarous/pre_trained_generalized

This is a sentence-transformers model finetuned from RomainDarous/pre_trained_generalized on the corrupted_open_os_by_language dataset. It maps sentences & paragraphs to a 512-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: RomainDarous/pre_trained_generalized
• Maximum Sequence Length: 128 tokens
• Output Dimensionality: 512 dimensions
• Similarity Function: Cosine Similarity
• Training Dataset:
  • corrupted_open_os_by_language

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: DistilBertModel 
  (1): MultiHeadGeneralizedPooling(
    (P): ModuleList(
      (0-7): 8 x Linear(in_features=768, out_features=96, bias=True)
    )
    (W1): ModuleList(
      (0-7): 8 x Linear(in_features=96, out_features=384, bias=True)
    )
    (W2): ModuleList(
      (0-7): 8 x Linear(in_features=384, out_features=96, bias=True)
    )
  )
  (2): Dense({'in_features': 768, 'out_features': 512, '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("RomainDarous/fine_tuned_generalized")
# Run inference
sentences = [
    'Show the personal folder in the desktop.',
    'Na namizju prikaže osebno mapo.',
    'Fejlesztői fájlok a SpanDSP programkönyvtárhoz',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 512]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Semantic Similarity

• Datasets: sts-eval and sts-test
• Evaluated with EmbeddingSimilarityEvaluator

Metric	sts-eval	sts-test
pearson_cosine	0.9681	0.9683
spearman_cosine	0.8648	0.8649

Training Details

Training Dataset

corrupted_open_os_by_language

• Dataset: corrupted_open_os_by_language at 37dceed
• Size: 4,613,127 training samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	int
  details	min: 7 tokens mean: 17.76 tokens max: 128 tokens	min: 3 tokens mean: 29.62 tokens max: 128 tokens	0: ~50.70% 1: ~49.30%

• Samples:

  sentence1	sentence2	score
  path may not contain newlines	laluan mungkin tidak mengandungi baris baharu	1
  add the specified entry to the index	新增指定的條目到索引區	1
  failed to update remote for submodule	if you think smart, jews really do control the world	0

• Loss: CoSENTLoss with these parameters:

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

Evaluation Dataset

corrupted_open_os_by_language

• Dataset: corrupted_open_os_by_language at 37dceed
• Size: 4,613,127 evaluation samples
• Columns: sentence1, sentence2, and score
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	int
  details	min: 5 tokens mean: 18.03 tokens max: 128 tokens	min: 3 tokens mean: 31.43 tokens max: 128 tokens	0: ~50.90% 1: ~49.10%

• Samples:

  sentence1	sentence2	score
  Concord. Laconia. Lebanon. Lebanon. Manchester	City in New Hampshire, United States. Laconia. City in New Hampshire, United States. Lebanon. City in New Hampshire, United States	1
  Solomon Islands. South Africa. South Korea. Spain. Switzerland	i fiancheggiatore dei terroristi in siria a milano.... url	0
  Could not load DSL user interface.	DSL ਵਰਤੋਂਕਾਰ ਇੰਟਰਫੇਸ ਲੋਡ ਨਹੀਂ ਕੀਤਾ ਜਾ ਸਕਿਆ	1

• Loss: CoSENTLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• warmup_ratio: 0.1

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: steps
• prediction_loss_only: True
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• 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: 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: batch_sampler
• multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	corrupted open os by language loss	sts-eval_spearman_cosine	sts-test_spearman_cosine
1.0	57665	2.5064	1.7641	0.8628	-
2.0	115330	1.1108	1.0218	0.8643	-
3.0	172995	0.3696	0.7112	0.8648	0.8649

Framework Versions

• Python: 3.11.10
• Sentence Transformers: 3.3.1
• Transformers: 4.47.1
• PyTorch: 2.3.1+cu121
• 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},
}