codersan/validadted_e5smallStudent

SentenceTransformer based on intfloat/multilingual-e5-small

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-small. 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: intfloat/multilingual-e5-small
• Maximum Sequence Length: 512 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': 512, '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})
  (2): Normalize()
)

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("codersan/validadted_e5smallStudent")
# Run inference
sentences = [
    'داشتن هزاران دنبال کننده در Quora چگونه است؟',
    'چه چیزی است که ده ها هزار دنبال کننده در Quora داشته باشید؟',
    'چگونه Airprint HP OfficeJet 4620 با HP LaserJet Enterprise M606X مقایسه می شود؟',
]
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]

Training Details

Training Dataset

Unnamed Dataset

• Size: 172,826 training 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: 16.19 tokens max: 84 tokens	min: 6 tokens mean: 16.5 tokens max: 52 tokens	min: 0.73 mean: 0.94 max: 1.0

• Samples:

  sentence1	sentence2	score
  تفاوت بین تحلیلگر تحقیقات بازار و تحلیلگر تجارت چیست؟	تفاوت بین تحقیقات بازاریابی و تحلیلگر تجارت چیست؟	0.9806554317474365
  خوردن چه چیزی باعث دل درد میشود؟	چه چیزی باعث رفع دل درد میشود؟	0.9417070150375366
  بهترین نرم افزار ویرایش ویدیویی کدام است؟	بهترین نرم افزار برای ویرایش ویدیو چیست؟	0.9928616285324097

• Loss: CosineSimilarityLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• eval_strategy: steps
• per_device_train_batch_size: 12
• learning_rate: 5e-06
• weight_decay: 0.01
• num_train_epochs: 1
• warmup_ratio: 0.1
• push_to_hub: True
• hub_model_id: codersan/validadted_e5smallStudent
• eval_on_start: True
• 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: 12
• 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: 5e-06
• weight_decay: 0.01
• adam_beta1: 0.9
• adam_beta2: 0.999
• adam_epsilon: 1e-08
• max_grad_norm: 1
• num_train_epochs: 1
• 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: True
• resume_from_checkpoint: None
• hub_model_id: codersan/validadted_e5smallStudent
• 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: True
• 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

Click to expand

Epoch	Step	Training Loss
0	0	-
0.0069	100	0.0004
0.0139	200	0.0004
0.0208	300	0.0003
0.0278	400	0.0003
0.0347	500	0.0003
0.0417	600	0.0003
0.0486	700	0.0003
0.0555	800	0.0003
0.0625	900	0.0003
0.0694	1000	0.0003
0.0764	1100	0.0002
0.0833	1200	0.0002
0.0903	1300	0.0002
0.0972	1400	0.0002
0.1041	1500	0.0002
0.1111	1600	0.0002
0.1180	1700	0.0002
0.1250	1800	0.0002
0.1319	1900	0.0002
0.1389	2000	0.0002
0.1458	2100	0.0002
0.1527	2200	0.0002
0.1597	2300	0.0002
0.1666	2400	0.0002
0.1736	2500	0.0002
0.1805	2600	0.0002
0.1875	2700	0.0002
0.1944	2800	0.0002
0.2013	2900	0.0002
0.2083	3000	0.0002
0.2152	3100	0.0002
0.2222	3200	0.0002
0.2291	3300	0.0002
0.2361	3400	0.0002
0.2430	3500	0.0002
0.2499	3600	0.0002
0.2569	3700	0.0002
0.2638	3800	0.0002
0.2708	3900	0.0002
0.2777	4000	0.0002
0.2847	4100	0.0002
0.2916	4200	0.0002
0.2985	4300	0.0002
0.3055	4400	0.0002
0.3124	4500	0.0002
0.3194	4600	0.0002
0.3263	4700	0.0002
0.3333	4800	0.0002
0.3402	4900	0.0002
0.3471	5000	0.0002
0.3541	5100	0.0002
0.3610	5200	0.0002
0.3680	5300	0.0002
0.3749	5400	0.0002
0.3819	5500	0.0002
0.3888	5600	0.0002
0.3958	5700	0.0002
0.4027	5800	0.0002
0.4096	5900	0.0002
0.4166	6000	0.0002
0.4235	6100	0.0002
0.4305	6200	0.0002
0.4374	6300	0.0002
0.4444	6400	0.0002
0.4513	6500	0.0002
0.4582	6600	0.0002
0.4652	6700	0.0002
0.4721	6800	0.0002
0.4791	6900	0.0002
0.4860	7000	0.0002
0.4930	7100	0.0002
0.4999	7200	0.0002
0.5068	7300	0.0002
0.5138	7400	0.0002
0.5207	7500	0.0002
0.5277	7600	0.0002
0.5346	7700	0.0002
0.5416	7800	0.0002
0.5485	7900	0.0002
0.5554	8000	0.0002
0.5624	8100	0.0002
0.5693	8200	0.0002
0.5763	8300	0.0002
0.5832	8400	0.0002
0.5902	8500	0.0002
0.5971	8600	0.0002
0.6040	8700	0.0002
0.6110	8800	0.0002
0.6179	8900	0.0002
0.6249	9000	0.0002
0.6318	9100	0.0002
0.6388	9200	0.0002
0.6457	9300	0.0002
0.6526	9400	0.0002
0.6596	9500	0.0002
0.6665	9600	0.0002
0.6735	9700	0.0002
0.6804	9800	0.0002
0.6874	9900	0.0002
0.6943	10000	0.0002
0.7012	10100	0.0002
0.7082	10200	0.0002
0.7151	10300	0.0002
0.7221	10400	0.0002
0.7290	10500	0.0002
0.7360	10600	0.0002
0.7429	10700	0.0002
0.7498	10800	0.0002
0.7568	10900	0.0002
0.7637	11000	0.0002
0.7707	11100	0.0002
0.7776	11200	0.0002
0.7846	11300	0.0002
0.7915	11400	0.0002
0.7984	11500	0.0002
0.8054	11600	0.0002
0.8123	11700	0.0002
0.8193	11800	0.0002
0.8262	11900	0.0002
0.8332	12000	0.0002
0.8401	12100	0.0002
0.8470	12200	0.0002
0.8540	12300	0.0002
0.8609	12400	0.0002
0.8679	12500	0.0002
0.8748	12600	0.0002
0.8818	12700	0.0002
0.8887	12800	0.0002
0.8956	12900	0.0002
0.9026	13000	0.0002
0.9095	13100	0.0002
0.9165	13200	0.0002
0.9234	13300	0.0002
0.9304	13400	0.0002
0.9373	13500	0.0002
0.9442	13600	0.0002
0.9512	13700	0.0002
0.9581	13800	0.0002
0.9651	13900	0.0002
0.9720	14000	0.0002
0.9790	14100	0.0002
0.9859	14200	0.0002
0.9928	14300	0.0002
0.9998	14400	0.0002

Framework Versions

• Python: 3.10.12
• Sentence Transformers: 3.3.1
• Transformers: 4.47.0
• PyTorch: 2.5.1+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",
}