Hugging Face's logo

srikarvar
/
fine_tuned_model_5

Sentence Similarity
sentence-transformers
Safetensors
bert
feature-extraction
Generated from Trainer
dataset_size:2332
loss:OnlineContrastiveLoss
Eval Results
text-embeddings-inference
Inference Endpoints
Model card Files Community
fine_tuned_model_5 / README.md
srikarvar's picture
srikarvar
Add new SentenceTransformer model.
4e4dc22 verified
preview code
|
raw
history blame contribute delete
26.5 kB
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:2332
  - loss:OnlineContrastiveLoss
widget:
  - source_sentence: Who discovered the structure of DNA?
    sentences:
      - Who discovered the structure of RNA?
      - Steps to apply for a scholarship
      - First human to set foot on the moon
  - source_sentence: Who directed 'Schindler's List'?
    sentences:
      - Who directed 'The Pianist'?
      - >-
        What are some high paying jobs for a fresher with an M.Tech in
        biotechnology?
      - Where can I find gluten-free restaurants?
  - source_sentence: Which is the best shares to purchase and sale daily trading?
    sentences:
      - >-
        In Sydney, which company would be the best to get advice for Business
        Sales & Purchases?
      - Steps to adjust phone settings
      - Is a 3.8 GPA sufficient to get into a top school?
  - source_sentence: Nd she is always sad?
    sentences:
      - Where to purchase organic produce
      - Aerodynamically what happens when propellor rotates?
      - How is unsupervised learning used for data insights?
  - source_sentence: How to bake a pie?
    sentences:
      - What is the population of Chicago?
      - Steps to bake a pie
      - >-
        What is the distribution of traffic between Google organic search
        results? e.g. #1 vs. #2 in rankings, first page vs. second page
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: 0.8653846153846154
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.872760534286499
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.8656716417910447
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.8200240135192871
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.8285714285714286
            name: Cosine Precision
          - type: cosine_recall
            value: 0.90625
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9322624848213654
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.8653846153846154
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 0.872760534286499
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.8656716417910447
            name: Dot F1
          - type: dot_f1_threshold
            value: 0.8200240135192871
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.8285714285714286
            name: Dot Precision
          - type: dot_recall
            value: 0.90625
            name: Dot Recall
          - type: dot_ap
            value: 0.9322624848213654
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.8692307692307693
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 9.252302169799805
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.8721804511278196
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 9.252302169799805
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.8405797101449275
            name: Manhattan Precision
          - type: manhattan_recall
            value: 0.90625
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9322911488571455
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.8653846153846154
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 0.5044240355491638
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.8656716417910447
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 0.5999571084976196
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.8285714285714286
            name: Euclidean Precision
          - type: euclidean_recall
            value: 0.90625
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.9322624848213654
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.8692307692307693
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 9.252302169799805
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.8721804511278196
            name: Max F1
          - type: max_f1_threshold
            value: 9.252302169799805
            name: Max F1 Threshold
          - type: max_precision
            value: 0.8405797101449275
            name: Max Precision
          - type: max_recall
            value: 0.90625
            name: Max Recall
          - type: max_ap
            value: 0.9322911488571455
            name: Max Ap
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: pair class test
          type: pair-class-test
        metrics:
          - type: cosine_accuracy
            value: 0.916
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.844039261341095
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.907488986784141
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.8230063319206238
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.8728813559322034
            name: Cosine Precision
          - type: cosine_recall
            value: 0.944954128440367
            name: Cosine Recall
          - type: cosine_ap
            value: 0.96095333014952
            name: Cosine Ap
          - type: dot_accuracy
            value: 0.916
            name: Dot Accuracy
          - type: dot_accuracy_threshold
            value: 0.8440393209457397
            name: Dot Accuracy Threshold
          - type: dot_f1
            value: 0.907488986784141
            name: Dot F1
          - type: dot_f1_threshold
            value: 0.8230063319206238
            name: Dot F1 Threshold
          - type: dot_precision
            value: 0.8728813559322034
            name: Dot Precision
          - type: dot_recall
            value: 0.944954128440367
            name: Dot Recall
          - type: dot_ap
            value: 0.96095333014952
            name: Dot Ap
          - type: manhattan_accuracy
            value: 0.916
            name: Manhattan Accuracy
          - type: manhattan_accuracy_threshold
            value: 8.581160545349121
            name: Manhattan Accuracy Threshold
          - type: manhattan_f1
            value: 0.907488986784141
            name: Manhattan F1
          - type: manhattan_f1_threshold
            value: 9.327116012573242
            name: Manhattan F1 Threshold
          - type: manhattan_precision
            value: 0.8728813559322034
            name: Manhattan Precision
          - type: manhattan_recall
            value: 0.944954128440367
            name: Manhattan Recall
          - type: manhattan_ap
            value: 0.9612698712458685
            name: Manhattan Ap
          - type: euclidean_accuracy
            value: 0.916
            name: Euclidean Accuracy
          - type: euclidean_accuracy_threshold
            value: 0.5584936141967773
            name: Euclidean Accuracy Threshold
          - type: euclidean_f1
            value: 0.907488986784141
            name: Euclidean F1
          - type: euclidean_f1_threshold
            value: 0.594968318939209
            name: Euclidean F1 Threshold
          - type: euclidean_precision
            value: 0.8728813559322034
            name: Euclidean Precision
          - type: euclidean_recall
            value: 0.944954128440367
            name: Euclidean Recall
          - type: euclidean_ap
            value: 0.96095333014952
            name: Euclidean Ap
          - type: max_accuracy
            value: 0.916
            name: Max Accuracy
          - type: max_accuracy_threshold
            value: 8.581160545349121
            name: Max Accuracy Threshold
          - type: max_f1
            value: 0.907488986784141
            name: Max F1
          - type: max_f1_threshold
            value: 9.327116012573242
            name: Max F1 Threshold
          - type: max_precision
            value: 0.8728813559322034
            name: Max Precision
          - type: max_recall
            value: 0.944954128440367
            name: Max Recall
          - type: max_ap
            value: 0.9612698712458685
            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/fine_tuned_model_5")
# Run inference
sentences = [
    'How to bake a pie?',
    'Steps to bake a pie',
    'What is the population of Chicago?',
]
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 0.8654
cosine_accuracy_threshold 0.8728
cosine_f1 0.8657
cosine_f1_threshold 0.82
cosine_precision 0.8286
cosine_recall 0.9062
cosine_ap 0.9323
dot_accuracy 0.8654
dot_accuracy_threshold 0.8728
dot_f1 0.8657
dot_f1_threshold 0.82
dot_precision 0.8286
dot_recall 0.9062
dot_ap 0.9323
manhattan_accuracy 0.8692
manhattan_accuracy_threshold 9.2523
manhattan_f1 0.8722
manhattan_f1_threshold 9.2523
manhattan_precision 0.8406
manhattan_recall 0.9062
manhattan_ap 0.9323
euclidean_accuracy 0.8654
euclidean_accuracy_threshold 0.5044
euclidean_f1 0.8657
euclidean_f1_threshold 0.6
euclidean_precision 0.8286
euclidean_recall 0.9062
euclidean_ap 0.9323
max_accuracy 0.8692
max_accuracy_threshold 9.2523
max_f1 0.8722
max_f1_threshold 9.2523
max_precision 0.8406
max_recall 0.9062
max_ap 0.9323

Binary Classification

Metric Value
cosine_accuracy 0.916
cosine_accuracy_threshold 0.844
cosine_f1 0.9075
cosine_f1_threshold 0.823
cosine_precision 0.8729
cosine_recall 0.945
cosine_ap 0.961
dot_accuracy 0.916
dot_accuracy_threshold 0.844
dot_f1 0.9075
dot_f1_threshold 0.823
dot_precision 0.8729
dot_recall 0.945
dot_ap 0.961
manhattan_accuracy 0.916
manhattan_accuracy_threshold 8.5812
manhattan_f1 0.9075
manhattan_f1_threshold 9.3271
manhattan_precision 0.8729
manhattan_recall 0.945
manhattan_ap 0.9613
euclidean_accuracy 0.916
euclidean_accuracy_threshold 0.5585
euclidean_f1 0.9075
euclidean_f1_threshold 0.595
euclidean_precision 0.8729
euclidean_recall 0.945
euclidean_ap 0.961
max_accuracy 0.916
max_accuracy_threshold 8.5812
max_f1 0.9075
max_f1_threshold 9.3271
max_precision 0.8729
max_recall 0.945
max_ap 0.9613

Training Details

Training Dataset

Unnamed Dataset

  • Size: 2,332 training samples
  • Columns: sentence1, sentence2, and label
  • Approximate statistics based on the first 1000 samples:
    sentence1 sentence2 label
    type string string int
    details
    • min: 6 tokens
    • mean: 12.96 tokens
    • max: 66 tokens
    • min: 5 tokens
    • mean: 12.67 tokens
    • max: 55 tokens
    • 0: ~52.80%
    • 1: ~47.20%
  • Samples:
    sentence1 sentence2 label
    How to bake a chocolate cake? Recipe for baking a chocolate cake 1
    Why do girls want to be friends with the guy they reject? How do guys feel after rejecting a girl? 0
    How can I stop being afraid of working? How do you stop being afraid of everything? 0
  • Loss: OnlineContrastiveLoss

Evaluation Dataset

Unnamed Dataset

  • Size: 260 evaluation samples
  • Columns: sentence1, sentence2, and label
  • Approximate statistics based on the first 1000 samples:
    sentence1 sentence2 label
    type string string int
    details
    • min: 6 tokens
    • mean: 13.44 tokens
    • max: 39 tokens
    • min: 4 tokens
    • mean: 12.99 tokens
    • max: 50 tokens
    • 0: ~50.77%
    • 1: ~49.23%
  • Samples:
    sentence1 sentence2 label
    How to cook spaghetti? Steps to cook spaghetti 1
    How to create a mobile app? How to create a desktop application? 0
    How can I update my resume? Steps to revise and update a resume 1
  • Loss: OnlineContrastiveLoss

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: epoch
  • per_device_train_batch_size: 32
  • per_device_eval_batch_size: 32
  • gradient_accumulation_steps: 2
  • num_train_epochs: 4
  • 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: 32
  • per_device_eval_batch_size: 32
  • per_gpu_train_batch_size: None
  • per_gpu_eval_batch_size: None
  • gradient_accumulation_steps: 2
  • eval_accumulation_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: 4
  • 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: 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 - - 0.6979 -
0.2740 10 1.9007 - - -
0.5479 20 1.1616 - - -
0.8219 30 0.9094 - - -
0.9863 36 - 0.7692 0.9117 -
1.0959 40 0.9105 - - -
1.3699 50 0.6629 - - -
1.6438 60 0.4243 - - -
1.9178 70 0.4729 - - -
2.0 73 - 0.7294 0.9306 -
2.1918 80 0.4897 - - -
2.4658 90 0.3103 - - -
2.7397 100 0.2316 - - -
2.9863 109 - 0.7807 0.9311 -
3.0137 110 0.3179 - - -
3.2877 120 0.1975 - - -
3.5616 130 0.1477 - - -
3.8356 140 0.1034 - - -
3.9452 144 - 0.8132 0.9323 0.9613
  • 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",
}