Husain's picture
Add new SentenceTransformer model.
31a3b85 verified
|
raw
history blame
20.1 kB
metadata
base_model: sentence-transformers/all-MiniLM-L6-v2
datasets:
  - sentence-transformers/stsb
language:
  - en
library_name: sentence-transformers
metrics:
  - pearson_cosine
  - spearman_cosine
  - pearson_manhattan
  - spearman_manhattan
  - pearson_euclidean
  - spearman_euclidean
  - pearson_dot
  - spearman_dot
  - pearson_max
  - spearman_max
pipeline_tag: sentence-similarity
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:101
  - loss:CoSENTLoss
widget:
  - source_sentence: The man is slicing a potato.
    sentences:
      - A woman is slicing carrot.
      - Two women are singing.
      - A man is slicing potato.
  - source_sentence: A girl is playing a flute.
    sentences:
      - A woman stirs eggs in a bowl.
      - A girl plays a wind instrument.
      - A man is turning over tables in anger.
  - source_sentence: People are playing baseball.
    sentences:
      - The cricket player hit the ball.
      - A man breaks a stick.
      - A woman is pouring a yellow mixture on a frying pan.
  - source_sentence: A woman and man are riding in a car.
    sentences:
      - A woman driving a car is talking to the man seated beside her.
      - A woman is placing skewered food onto a cooker.
      - The man and woman are walking.
  - source_sentence: A cat is on a robot.
    sentences:
      - A man is eating bread.
      - A woman is pouring eyes into a bowl.
      - A boy sits on a bed, sings and plays a guitar.
model-index:
  - name: SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2
    results:
      - task:
          type: semantic-similarity
          name: Semantic Similarity
        dataset:
          name: sts dev
          type: sts-dev
        metrics:
          - type: pearson_cosine
            value: 0.9186522039312566
            name: Pearson Cosine
          - type: spearman_cosine
            value: 0.9276278198564623
            name: Spearman Cosine
          - type: pearson_manhattan
            value: 0.8991493568260668
            name: Pearson Manhattan
          - type: spearman_manhattan
            value: 0.9320766471557739
            name: Spearman Manhattan
          - type: pearson_euclidean
            value: 0.9014580823459483
            name: Pearson Euclidean
          - type: spearman_euclidean
            value: 0.9289530024562572
            name: Spearman Euclidean
          - type: pearson_dot
            value: 0.8789190604301875
            name: Pearson Dot
          - type: spearman_dot
            value: 0.8957287815613981
            name: Spearman Dot
          - type: pearson_max
            value: 0.9186522039312566
            name: Pearson Max
          - type: spearman_max
            value: 0.9320766471557739
            name: Spearman Max

SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2. 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: sentence-transformers/all-MiniLM-L6-v2
  • Maximum Sequence Length: 512 tokens
  • Output Dimensionality: 384 tokens
  • Similarity Function: Cosine Similarity
  • Language: en

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})
)

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("Husain/ramdam_fingerprint_embedding_model")
# Run inference
sentences = [
    'A cat is on a robot.',
    'A man is eating bread.',
    'A woman is pouring eyes into a bowl.',
]
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

Semantic Similarity

Metric Value
pearson_cosine 0.9187
spearman_cosine 0.9276
pearson_manhattan 0.8991
spearman_manhattan 0.9321
pearson_euclidean 0.9015
spearman_euclidean 0.929
pearson_dot 0.8789
spearman_dot 0.8957
pearson_max 0.9187
spearman_max 0.9321

Training Details

Training Dataset

Unnamed Dataset

  • Size: 101 training samples
  • Columns: sentence1, sentence2, and score
  • Approximate statistics based on the first 101 samples:
    sentence1 sentence2 score
    type string string float
    details
    • min: 7 tokens
    • mean: 9.44 tokens
    • max: 14 tokens
    • min: 3 tokens
    • mean: 9.46 tokens
    • max: 15 tokens
    • min: 0.1
    • mean: 0.66
    • max: 1.0
  • Samples:
    sentence1 sentence2 score
    A plane is taking off. An air plane is taking off. 1.0
    A man is playing a large flute. A man is playing a flute. 0.76
    A man is spreading shreded cheese on a pizza. A man is spreading shredded cheese on an uncooked pizza. 0.76
  • Loss: CoSENTLoss with these parameters:
    {
        "scale": 20.0,
        "similarity_fct": "pairwise_cos_sim"
    }
    

Evaluation Dataset

stsb

  • Dataset: stsb at ab7a5ac
  • Size: 1,500 evaluation 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: 9.35 tokens
    • max: 13 tokens
    • min: 7 tokens
    • mean: 9.9 tokens
    • max: 16 tokens
    • min: 0.0
    • mean: 0.39
    • max: 1.0
  • Samples:
    sentence1 sentence2 score
    A woman is riding on a horse. A man is turning over tables in anger. 0.0
    A man is screwing wood to a wall. A man is giving a woman a massage. 0.04
    A girl is playing a flute. A girl plays a wind instrument. 0.64
  • Loss: CoSENTLoss with these parameters:
    {
        "scale": 20.0,
        "similarity_fct": "pairwise_cos_sim"
    }
    

Training Hyperparameters

Non-Default Hyperparameters

  • eval_strategy: steps
  • learning_rate: 2e-05
  • num_train_epochs: 10
  • warmup_ratio: 0.1
  • save_only_model: True
  • seed: 33
  • fp16: True
  • load_best_model_at_end: True

All Hyperparameters

Click to expand
  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: steps
  • prediction_loss_only: True
  • per_device_train_batch_size: 8
  • 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: 0.0
  • adam_beta1: 0.9
  • adam_beta2: 0.999
  • adam_epsilon: 1e-08
  • max_grad_norm: 1.0
  • num_train_epochs: 10
  • 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: True
  • restore_callback_states_from_checkpoint: False
  • no_cuda: False
  • use_cpu: False
  • use_mps_device: False
  • seed: 33
  • data_seed: None
  • jit_mode_eval: False
  • use_ipex: False
  • bf16: False
  • fp16: True
  • 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
  • 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
  • eval_on_start: False
  • eval_use_gather_object: False
  • batch_sampler: batch_sampler
  • multi_dataset_batch_sampler: proportional

Training Logs

Epoch Step loss sts-dev_spearman_cosine
0.1538 2 4.4641 0.9366
0.3077 4 4.4652 0.9366
0.4615 6 4.4719 0.9366
0.6154 8 4.4903 0.9366
0.7692 10 4.5264 0.9373
0.9231 12 4.5954 0.9339
1.0769 14 4.6832 0.9328
1.2308 16 4.7534 0.9289
1.3846 18 4.8155 0.9281
1.5385 20 4.8788 0.9269
1.6923 22 4.9350 0.9272
1.8462 24 4.9789 0.9239
2.0 26 5.0132 0.9230
2.1538 28 5.0636 0.9237
2.3077 30 5.1068 0.9202
2.4615 32 5.1460 0.9172
2.6154 34 5.1602 0.9164
2.7692 36 5.1493 0.9210
2.9231 38 5.1399 0.9200
3.0769 40 5.1342 0.9235
3.2308 42 5.1413 0.9258
3.3846 44 5.1440 0.9271
3.5385 46 5.1583 0.9311
3.6923 48 5.1664 0.9293
3.8462 50 5.1682 0.9293
4.0 52 5.1617 0.9293
4.1538 54 5.1543 0.9293
4.3077 56 5.1480 0.9293
4.4615 58 5.1428 0.9291
4.6154 60 5.1292 0.9298
4.7692 62 5.1271 0.9276
4.9231 64 5.1133 0.9276
5.0769 66 5.0928 0.9270
5.2308 68 5.0874 0.9270
5.3846 70 5.0755 0.9270
5.5385 72 5.0665 0.9270
5.6923 74 5.0676 0.9293
5.8462 76 5.0747 0.9293
6.0 78 5.0647 0.9295
6.1538 80 5.0763 0.9273
6.3077 82 5.0832 0.9272
6.4615 84 5.0750 0.9289
6.6154 86 5.0547 0.9289
6.7692 88 5.0350 0.9308
6.9231 90 5.0221 0.9308
7.0769 92 5.0107 0.9308
7.2308 94 4.9967 0.9297
7.3846 96 4.9983 0.9297
7.5385 98 5.0026 0.9277
7.6923 100 5.0095 0.9277
7.8462 102 5.0102 0.9277
8.0 104 5.0055 0.9271
8.1538 106 5.0031 0.9271
8.3077 108 4.9976 0.9271
8.4615 110 4.9941 0.9271
8.6154 112 4.9856 0.9276
8.7692 114 4.9821 0.9276
8.9231 116 4.9782 0.9276
9.0769 118 4.9706 0.9276
9.2308 120 4.9646 0.9276
9.3846 122 4.9584 0.9276
9.5385 124 4.9537 0.9276
9.6923 126 4.9499 0.9276
9.8462 128 4.9485 0.9276
10.0 130 4.9463 0.9276
  • The bold row denotes the saved checkpoint.

Framework Versions

  • Python: 3.8.10
  • Sentence Transformers: 3.1.0
  • Transformers: 4.44.2
  • PyTorch: 2.3.1+cu121
  • Accelerate: 0.34.2
  • Datasets: 3.0.0
  • 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",
}

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},
}