codersan/newfa_e5base2

SentenceTransformer based on intfloat/multilingual-e5-base

This is a sentence-transformers model finetuned from intfloat/multilingual-e5-base. 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: intfloat/multilingual-e5-base
• 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: XLMRobertaModel 
  (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): 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/newfa_e5base2")
# Run inference
sentences = [
    'مرزهای صفحه چیست؟برخی از انواع چیست؟',
    'مرزهای صفحه چیست؟',
    'اتانول چند ایزومر دارد؟',
]
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]

Training Details

Training Dataset

Unnamed Dataset

• Size: 142,964 training samples
• Columns: anchor and positive
• Approximate statistics based on the first 1000 samples:

  	anchor	positive
  type	string	string
  details	min: 6 tokens mean: 16.39 tokens max: 90 tokens	min: 6 tokens mean: 15.68 tokens max: 57 tokens

• Samples:

  anchor	positive
  گاو یونجه می خورد	گاو در حال چریدن است
  ماشینی به شکلی خطرناک از روی دختری می‌پرد.	دختر با بی‌احتیاطی روی ماشین می‌پرد.
  چگونه می توانم کارتهای هدیه iTunes رایگان را در هند دریافت کنم؟	چگونه می توانم کارتهای هدیه iTunes رایگان دریافت کنم؟

• Loss: MultipleNegativesRankingLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 32
• learning_rate: 2e-05
• weight_decay: 0.01
• batch_sampler: no_duplicates

All Hyperparameters

Click to expand

• overwrite_output_dir: False
• do_predict: False
• eval_strategy: no
• prediction_loss_only: True
• per_device_train_batch_size: 32
• 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: 0.01
• 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: 0.0
• 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

Click to expand

Epoch	Step	Training Loss
0.0224	100	0.0821
0.0448	200	0.0455
0.0671	300	0.0408
0.0895	400	0.0461
0.1119	500	0.0418
0.1343	600	0.0449
0.1567	700	0.0314
0.1791	800	0.0252
0.2014	900	0.0254
0.2238	1000	0.0341
0.2462	1100	0.0239
0.2686	1200	0.0308
0.2910	1300	0.0415
0.3133	1400	0.0386
0.3357	1500	0.027
0.3581	1600	0.0369
0.3805	1700	0.0346
0.4029	1800	0.0301
0.4252	1900	0.03
0.4476	2000	0.0179
0.4700	2100	0.035
0.4924	2200	0.0327
0.5148	2300	0.033
0.5372	2400	0.0272
0.5595	2500	0.0318
0.5819	2600	0.025
0.6043	2700	0.023
0.6267	2800	0.0294
0.6491	2900	0.0337
0.6714	3000	0.0274
0.6938	3100	0.0223
0.7162	3200	0.0384
0.7386	3300	0.0217
0.7610	3400	0.032
0.7833	3500	0.0309
0.8057	3600	0.024
0.8281	3700	0.0273
0.8505	3800	0.0245
0.8729	3900	0.0268
0.8953	4000	0.0322
0.9176	4100	0.0271
0.9400	4200	0.0316
0.9624	4300	0.0179
0.9848	4400	0.0294
1.0072	4500	0.0283
1.0295	4600	0.0171
1.0519	4700	0.017
1.0743	4800	0.0197
1.0967	4900	0.0215
1.1191	5000	0.02
1.1415	5100	0.0144
1.1638	5200	0.015
1.1862	5300	0.0084
1.2086	5400	0.0115
1.2310	5500	0.0143
1.2534	5600	0.0129
1.2757	5700	0.0165
1.2981	5800	0.0168
1.3205	5900	0.0233
1.3429	6000	0.0156
1.3653	6100	0.0207
1.3876	6200	0.0149
1.4100	6300	0.0134
1.4324	6400	0.0108
1.4548	6500	0.0118
1.4772	6600	0.0173
1.4996	6700	0.0171
1.5219	6800	0.0168
1.5443	6900	0.0144
1.5667	7000	0.0111
1.5891	7100	0.0117
1.6115	7200	0.0122
1.6338	7300	0.0143
1.6562	7400	0.0151
1.6786	7500	0.0152
1.7010	7600	0.012
1.7234	7700	0.0177
1.7457	7800	0.0172
1.7681	7900	0.016
1.7905	8000	0.0141
1.8129	8100	0.0112
1.8353	8200	0.011
1.8577	8300	0.0132
1.8800	8400	0.0127
1.9024	8500	0.0188
1.9248	8600	0.0196
1.9472	8700	0.0106
1.9696	8800	0.0108
1.9919	8900	0.0172
2.0143	9000	0.0116
2.0367	9100	0.0089
2.0591	9200	0.0096
2.0815	9300	0.0142
2.1038	9400	0.0112
2.1262	9500	0.0103
2.1486	9600	0.0077
2.1710	9700	0.0082
2.1934	9800	0.0066
2.2158	9900	0.0106
2.2381	10000	0.0072
2.2605	10100	0.0085
2.2829	10200	0.0085
2.3053	10300	0.015
2.3277	10400	0.0113
2.3500	10500	0.0118
2.3724	10600	0.0123
2.3948	10700	0.0071
2.4172	10800	0.0087
2.4396	10900	0.0056
2.4620	11000	0.0091
2.4843	11100	0.0116
2.5067	11200	0.0123
2.5291	11300	0.0108
2.5515	11400	0.0078
2.5739	11500	0.0072
2.5962	11600	0.0084
2.6186	11700	0.0066
2.6410	11800	0.0115
2.6634	11900	0.0088
2.6858	12000	0.008
2.7081	12100	0.0095
2.7305	12200	0.0108
2.7529	12300	0.0113
2.7753	12400	0.0086
2.7977	12500	0.0096
2.8201	12600	0.0093
2.8424	12700	0.0076
2.8648	12800	0.006
2.8872	12900	0.0124
2.9096	13000	0.0131
2.9320	13100	0.0103
2.9543	13200	0.0063
2.9767	13300	0.0067
2.9991	13400	0.0117

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: 4.0.0
• 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",
}

MultipleNegativesRankingLoss

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}