srikarvar's picture
Add new SentenceTransformer model.
8f7c78e verified
metadata
base_model: intfloat/multilingual-e5-small
datasets: []
language: []
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - cosine_accuracy_threshold
  - cosine_f1
  - cosine_f1_threshold
  - cosine_precision
  - cosine_recall
  - cosine_ap
  - dot_accuracy
  - dot_accuracy_threshold
  - dot_f1
  - dot_f1_threshold
  - dot_precision
  - dot_recall
  - dot_ap
  - manhattan_accuracy
  - manhattan_accuracy_threshold
  - manhattan_f1
  - manhattan_f1_threshold
  - manhattan_precision
  - manhattan_recall
  - manhattan_ap
  - euclidean_accuracy
  - euclidean_accuracy_threshold
  - euclidean_f1
  - euclidean_f1_threshold
  - euclidean_precision
  - euclidean_recall
  - euclidean_ap
  - max_accuracy
  - max_accuracy_threshold
  - max_f1
  - max_f1_threshold
  - max_precision
  - max_recall
  - max_ap
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:296
  - loss:ContrastiveLoss
widget:
  - source_sentence: Biography of Queen Elisabeth II
    sentences:
      - Biography of Queen Elisabeth I
      - What are the ingredients of a pizza?
      - When was the Declaration of Independence signed?
  - source_sentence: What are ingredients of pizza
    sentences:
      - How do I reset my password?
      - What are the ingredients of a pizza?
      - What's the capital of France?
  - source_sentence: Japan's population count
    sentences:
      - What's the population of Japan?
      - What is the time now?
      - What are the benefits of exercise?
  - source_sentence: How many states are in India?
    sentences:
      - What is the capital of Italy?
      - What is the speed of light?
      - How many states are in the USA?
  - source_sentence: Language spoken by the most people
    sentences:
      - What is the most spoken language in the world?
      - How do I get to the nearest train station?
      - Who was the first person to walk on the moon?
model-index:
  - name: SentenceTransformer based on intfloat/multilingual-e5-small
    results:
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: pair class dev
          type: pair-class-dev
        metrics:
          - type: cosine_accuracy
            value: 0.9459459459459459
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.8864079713821411
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.951219512195122
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.8166882395744324
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.9069767441860465
            name: Cosine Precision
          - type: cosine_recall
            value: 1
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9896978310027928
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.9459459459459459
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 0.8864079713821411
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.951219512195122
            name: Dot F1
          - type: dot_f1_threshold
            value: 0.8166882991790771
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.9069767441860465
            name: Dot Precision
          - type: dot_recall
            value: 1
            name: Dot Recall
          - type: dot_ap
            value: 0.9896978310027928
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.9459459459459459
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 7.30389404296875
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.951219512195122
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 9.542933464050293
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.9069767441860465
            name: Manhattan Precision
          - type: manhattan_recall
            value: 1
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9896978310027928
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.9459459459459459
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 0.4764680862426758
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.951219512195122
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 0.6043685078620911
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.9069767441860465
            name: Euclidean Precision
          - type: euclidean_recall
            value: 1
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.9896978310027928
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.9459459459459459
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 7.30389404296875
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.951219512195122
            name: Max F1
          - type: max_f1_threshold
            value: 9.542933464050293
            name: Max F1 Threshold
          - type: max_precision
            value: 0.9069767441860465
            name: Max Precision
          - type: max_recall
            value: 1
            name: Max Recall
          - type: max_ap
            value: 0.9896978310027928
            name: Max Ap
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: pair class test
          type: pair-class-test
        metrics:
          - type: cosine_accuracy
            value: 0.9459459459459459
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.8864079713821411
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.951219512195122
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.8166882395744324
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.9069767441860465
            name: Cosine Precision
          - type: cosine_recall
            value: 1
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9896978310027928
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.9459459459459459
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 0.8864079713821411
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.951219512195122
            name: Dot F1
          - type: dot_f1_threshold
            value: 0.8166882991790771
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.9069767441860465
            name: Dot Precision
          - type: dot_recall
            value: 1
            name: Dot Recall
          - type: dot_ap
            value: 0.9896978310027928
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.9459459459459459
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 7.30389404296875
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.951219512195122
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 9.542933464050293
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.9069767441860465
            name: Manhattan Precision
          - type: manhattan_recall
            value: 1
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9896978310027928
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.9459459459459459
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 0.4764680862426758
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.951219512195122
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 0.6043685078620911
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.9069767441860465
            name: Euclidean Precision
          - type: euclidean_recall
            value: 1
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.9896978310027928
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.9459459459459459
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 7.30389404296875
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.951219512195122
            name: Max F1
          - type: max_f1_threshold
            value: 9.542933464050293
            name: Max F1 Threshold
          - type: max_precision
            value: 0.9069767441860465
            name: Max Precision
          - type: max_recall
            value: 1
            name: Max Recall
          - type: max_ap
            value: 0.9896978310027928
            name: Max Ap

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 tokens
  • Similarity Function: Cosine Similarity

Model Sources

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("srikarvar/multilingual-e5-small-pairclass-contrastive")
# Run inference
sentences = [
    'Language spoken by the most people',
    'What is the most spoken language in the world?',
    'Who was the first person to walk on the moon?',
]
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

Metric Value
cosine_accuracy 0.9459
cosine_accuracy_threshold 0.8864
cosine_f1 0.9512
cosine_f1_threshold 0.8167
cosine_precision 0.907
cosine_recall 1.0
cosine_ap 0.9897
dot_accuracy 0.9459
dot_accuracy_threshold 0.8864
dot_f1 0.9512
dot_f1_threshold 0.8167
dot_precision 0.907
dot_recall 1.0
dot_ap 0.9897
manhattan_accuracy 0.9459
manhattan_accuracy_threshold 7.3039
manhattan_f1 0.9512
manhattan_f1_threshold 9.5429
manhattan_precision 0.907
manhattan_recall 1.0
manhattan_ap 0.9897
euclidean_accuracy 0.9459
euclidean_accuracy_threshold 0.4765
euclidean_f1 0.9512
euclidean_f1_threshold 0.6044
euclidean_precision 0.907
euclidean_recall 1.0
euclidean_ap 0.9897
max_accuracy 0.9459
max_accuracy_threshold 7.3039
max_f1 0.9512
max_f1_threshold 9.5429
max_precision 0.907
max_recall 1.0
max_ap 0.9897

Binary Classification

Metric Value
cosine_accuracy 0.9459
cosine_accuracy_threshold 0.8864
cosine_f1 0.9512
cosine_f1_threshold 0.8167
cosine_precision 0.907
cosine_recall 1.0
cosine_ap 0.9897
dot_accuracy 0.9459
dot_accuracy_threshold 0.8864
dot_f1 0.9512
dot_f1_threshold 0.8167
dot_precision 0.907
dot_recall 1.0
dot_ap 0.9897
manhattan_accuracy 0.9459
manhattan_accuracy_threshold 7.3039
manhattan_f1 0.9512
manhattan_f1_threshold 9.5429
manhattan_precision 0.907
manhattan_recall 1.0
manhattan_ap 0.9897
euclidean_accuracy 0.9459
euclidean_accuracy_threshold 0.4765
euclidean_f1 0.9512
euclidean_f1_threshold 0.6044
euclidean_precision 0.907
euclidean_recall 1.0
euclidean_ap 0.9897
max_accuracy 0.9459
max_accuracy_threshold 7.3039
max_f1 0.9512
max_f1_threshold 9.5429
max_precision 0.907
max_recall 1.0
max_ap 0.9897

Training Details

Training Dataset

Unnamed Dataset

  • Size: 296 training samples
  • Columns: label, sentence2, and sentence1
  • Approximate statistics based on the first 1000 samples:
    label sentence2 sentence1
    type int string string
    details
    • 0: ~50.68%
    • 1: ~49.32%
    • min: 4 tokens
    • mean: 9.39 tokens
    • max: 20 tokens
    • min: 6 tokens
    • mean: 10.24 tokens
    • max: 20 tokens
  • Samples:
    label sentence2 sentence1
    0 How to improve running speed? How to train for a marathon?
    0 What is the distance of a marathon? How to train for a marathon?
    1 Mona Lisa painter Who painted the Mona Lisa?
  • Loss: ContrastiveLoss with these parameters:
    {
        "distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
        "margin": 0.5,
        "size_average": true
    }
    

Evaluation Dataset

Unnamed Dataset

  • Size: 74 evaluation samples
  • Columns: label, sentence2, and sentence1
  • Approximate statistics based on the first 1000 samples:
    label sentence2 sentence1
    type int string string
    details
    • 0: ~47.30%
    • 1: ~52.70%
    • min: 5 tokens
    • mean: 9.18 tokens
    • max: 22 tokens
    • min: 7 tokens
    • mean: 10.15 tokens
    • max: 20 tokens
  • Samples:
    label sentence2 sentence1
    1 Bitcoin's current value What is the price of Bitcoin?
    1 Who found out about gravity? Who discovered gravity?
    1 Language spoken by the most people What is the most spoken language in the world?
  • Loss: ContrastiveLoss with these parameters:
    {
        "distance_metric": "SiameseDistanceMetric.COSINE_DISTANCE",
        "margin": 0.5,
        "size_average": true
    }
    

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: epoch
  • per_device_train_batch_size: 16
  • per_device_eval_batch_size: 16
  • gradient_accumulation_steps: 2
  • learning_rate: 3e-05
  • weight_decay: 0.01
  • num_train_epochs: 5
  • lr_scheduler_type: reduce_lr_on_plateau
  • warmup_ratio: 0.1
  • load_best_model_at_end: True
  • optim: adamw_torch_fused
  • batch_sampler: no_duplicates

All Hyperparameters

Click to expand
  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: epoch
  • 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: 2
  • eval_accumulation_steps: None
  • learning_rate: 3e-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: 5
  • max_steps: -1
  • lr_scheduler_type: reduce_lr_on_plateau
  • 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: True
  • 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_fused
  • 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: False
  • hub_always_push: False
  • gradient_checkpointing: False
  • gradient_checkpointing_kwargs: None
  • include_inputs_for_metrics: False
  • 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
  • batch_sampler: no_duplicates
  • multi_dataset_batch_sampler: proportional

Training Logs

Epoch Step Training Loss loss pair-class-dev_max_ap pair-class-test_max_ap
0 0 - - 0.6933 -
0.9474 9 - 0.0182 0.9142 -
1.0526 10 0.0311 - - -
2.0 19 - 0.0091 0.9730 -
2.1053 20 0.0119 - - -
2.9474 28 - 0.0090 0.9878 -
3.1579 30 0.0074 - - -
4.0 38 - 0.0084 0.9891 -
4.2105 40 0.005 - - -
4.7368 45 - 0.0084 0.9897 0.9897
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.10.12
  • Sentence Transformers: 3.0.1
  • Transformers: 4.41.2
  • PyTorch: 2.1.2+cu121
  • Accelerate: 0.32.1
  • Datasets: 2.19.1
  • Tokenizers: 0.19.1

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",
}

ContrastiveLoss

@inproceedings{hadsell2006dimensionality,
    author={Hadsell, R. and Chopra, S. and LeCun, Y.},
    booktitle={2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR'06)}, 
    title={Dimensionality Reduction by Learning an Invariant Mapping}, 
    year={2006},
    volume={2},
    number={},
    pages={1735-1742},
    doi={10.1109/CVPR.2006.100}
}