Add new SentenceTransformer model

2aae66d verified over 1 year ago

17.7 kB

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:84524
  - loss:AttributeTripletLoss
base_model: Alibaba-NLP/gte-base-en-v1.5
widget:
  - source_sentence: Don Piper
    sentences:
      - Tommy Nelson
      - Kate Walbert
      - publisher
      - author
  - source_sentence: The Luxe
    sentences:
      - '1999'
      - publication_date
      - title
      - 'Critical Care, Mercy Hospital Series #1'
  - source_sentence: Bram Stoker
    sentences:
      - author
      - Michael J. Pangio
      - '9781598871012'
      - isbn_13
  - source_sentence: '9780385340557'
    sentences:
      - BBC Books
      - '9780399208539'
      - author
      - isbn_13
  - source_sentence: Midnight
    sentences:
      - The Bone Parade
      - 12/01/2005
      - publication_date
      - title
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - silhouette_cosine
  - silhouette_euclidean
model-index:
  - name: SentenceTransformer based on Alibaba-NLP/gte-base-en-v1.5
    results:
      - task:
          type: triplet
          name: Triplet
        dataset:
          name: Unknown
          type: unknown
        metrics:
          - type: cosine_accuracy
            value: 0.1667376458644867
            name: Cosine Accuracy
          - type: cosine_accuracy
            value: 0.16471828520298004
            name: Cosine Accuracy
      - task:
          type: silhouette
          name: Silhouette
        dataset:
          name: Unknown
          type: unknown
        metrics:
          - type: silhouette_cosine
            value: -0.44084376096725464
            name: Silhouette Cosine
          - type: silhouette_euclidean
            value: 0.015225890092551708
            name: Silhouette Euclidean
          - type: silhouette_cosine
            value: -0.48077088594436646
            name: Silhouette Cosine
          - type: silhouette_euclidean
            value: 0.01669926382601261
            name: Silhouette Euclidean

SentenceTransformer based on Alibaba-NLP/gte-base-en-v1.5

This is a sentence-transformers model finetuned from Alibaba-NLP/gte-base-en-v1.5. 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: Alibaba-NLP/gte-base-en-v1.5
Maximum Sequence Length: 8192 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': 8192, 'do_lower_case': False}) with Transformer model: NewModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': True, 'pooling_mode_mean_tokens': False, '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("albertus-sussex/veriscrape-book-test-sbert-bs128_lr0.0002_ep3_cosine_snTrue_spFalse_hn1")
# Run inference
sentences = [
    'Midnight',
    'The Bone Parade',
    '12/01/2005',
]
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]

Evaluation

Metrics

Triplet

Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.1667

Silhouette

Evaluated with veriscrape.training.SilhouetteEvaluator

Metric	Value
silhouette_cosine	-0.4408
silhouette_euclidean	0.0152

Triplet

Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.1647

Silhouette

Evaluated with veriscrape.training.SilhouetteEvaluator

Metric	Value
silhouette_cosine	-0.4808
silhouette_euclidean	0.0167

Training Details

Training Dataset

Unnamed Dataset

Size: 84,524 training samples
Columns: anchor, positive, negative, pos_attr_name, and neg_attr_name

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative	pos_attr_name	neg_attr_name
type	string	string	string	string	string
details	min: 3 tokens mean: 6.97 tokens max: 37 tokens	min: 3 tokens mean: 7.09 tokens max: 28 tokens	min: 3 tokens mean: 6.31 tokens max: 23 tokens	min: 3 tokens mean: 3.77 tokens max: 5 tokens	min: 3 tokens mean: 3.8 tokens max: 5 tokens

Samples:

anchor	positive	negative	pos_attr_name	neg_attr_name
`09/01/1997`	`12/01/1977`	`2010`	`publication_date`	`title`
`9780060275730`	`9780829748772`	`HarperCollins Publishers Ltd`	`isbn_13`	`publisher`
`9780609809648`	`9780764551956`	`HarperCollins Publishers`	`isbn_13`	`author`

Loss: veriscrape.training.AttributeTripletLoss with these parameters:

{
    "distance_metric": "TripletDistanceMetric.COSINE",
    "triplet_margin": 5
}

Evaluation Dataset

Unnamed Dataset

Size: 9,392 evaluation samples
Columns: anchor, positive, negative, pos_attr_name, and neg_attr_name

Approximate statistics based on the first 1000 samples:

	anchor	positive	negative	pos_attr_name	neg_attr_name
type	string	string	string	string	string
details	min: 3 tokens mean: 6.85 tokens max: 27 tokens	min: 3 tokens mean: 6.98 tokens max: 44 tokens	min: 3 tokens mean: 6.08 tokens max: 18 tokens	min: 3 tokens mean: 3.75 tokens max: 5 tokens	min: 3 tokens mean: 3.8 tokens max: 5 tokens

Samples:

anchor	positive	negative	pos_attr_name	neg_attr_name
`9780764200564`	`: 9780590458467`	`1984`	`isbn_13`	`publication_date`
`Penguin Group USA`	`Signet`	`9781600243912`	`publisher`	`isbn_13`
`Alphabet Juice`	`Space`	`9780807871133`	`title`	`isbn_13`

Loss: veriscrape.training.AttributeTripletLoss with these parameters:

{
    "distance_metric": "TripletDistanceMetric.COSINE",
    "triplet_margin": 5
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 128
per_device_eval_batch_size: 128
learning_rate: 0.0002
warmup_ratio: 0.1

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 128
per_device_eval_batch_size: 128
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: 0.0002
weight_decay: 0.0
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: 0.1
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: 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
use_liger_kernel: False
eval_use_gather_object: False
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss	cosine_accuracy	silhouette_cosine
-1	-1	-	-	0.4284	0.1492
1.0	661	4.0498	4.3099	0.6033	-0.2282
2.0	1322	4.4641	5.0	0.0400	0.0
3.0	1983	4.9651	5.0	0.1667	-0.4408
-1	-1	-	-	0.1647	-0.4808

Framework Versions

Python: 3.10.16
Sentence Transformers: 3.4.1
Transformers: 4.45.2
PyTorch: 2.5.1+cu124
Accelerate: 1.5.2
Datasets: 3.1.0
Tokenizers: 0.20.3

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",
}

AttributeTripletLoss

@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification},
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}