metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:2000
  - loss:CoSENTLoss
base_model: avsolatorio/GIST-small-Embedding-v0
widget:
  - source_sentence: is alexa compatible with tv?
    sentences:
      - >-
        Of een ei iedere dag gezond of ongezond is, hangt af van wat je verder
        iedere dag eet. Het Voedingscentrum adviseert om te variëren in vis,
        peulvruchten, vlees en ei. Het eten van 2-3 eieren per week past in een
        gezonde voeding. Vegetariërs kunnen 3-4 eieren per week eten.
      - >-
        The price was right, the size was right and as it turns out this PYLE TV
        has the best picture quality of all 5 TVs that our family watches! The
        setup was super easy with no hassle. I would recommend it to anyone!
      - >-
        According to the Association of British Insurers, insurance companies
        will look into a policyholder's medical profile if they give up smoking.
        They'll commonly seek a report from a policyholder's family doctor. If
        this raises concerns, they may ask a policyholder to have a chest X-ray.
  - source_sentence: is nyada a real college?
    sentences:
      - >-
        The instruments have been classified as Wind instruments (aero phonic)
        including Bansuri and Nagaswaram; String instruments (chordophonic)
        including Dilruba and Veena; Percussion instruments (membranophonic)
        including Tabla, Mridangam and (idiophonic) Bortal, and Ghatam.
      - >-
        This service is currently offered free of charge by the bank. You can
        get the last 'Available' balance of your account (by an SMS) by giving a
        Missed Call to 18008431122. You can get the Mini Statement (by an SMS)
        for last 5 transactions in your account by giving a Missed Call to
        18008431133. 1.
      - >-
        King Size Bed Known as a standard 5ft bed or 150cm wide by 200cm in
        length.
  - source_sentence: is europe bigger than australia?
    sentences:
      - >-
        Although this is just five per cent of the world's land mass (149.45
        million square kilometres), Australia is the planet's sixth largest
        country after Russia, Canada, China, the United States of America and
        Brazil. ... almost as great as that of the United States of America.
        about 50 per cent greater than Europe, and.
      - >-
        The recommended dose of evening primrose oil is 8 to 12 capsules a day,
        at a dose of 500 milligrams per capsule. A range of evening primrose oil
        products are available for purchase online.
      - >-
        This includes a three-year law degree, a one-year LPC and finally a
        two-year training contract with a law firm. Studying a non-law subject
        for your degree means you'll need to take the GDL conversion course
        before your LPC, which adds one year to the total.
  - source_sentence: how long does money take to transfer boi?
    sentences:
      - >-
        When will it take more than one working day? It will take more than one
        working day to reach your payee's bank when: You make a payment online
        after 3.30pm in the Republic of Ireland or after 4.30pm in Northern
        Ireland and Great Britain on a working day. Your payment will begin to
        process on the next working day.
      - >-
        U.S. citizens travelling to South Korea for business or tourism do not
        need a visa. ... Although obtaining a visa in advance can ease the entry
        process, as long as you have a valid U.S. passport, you can enter the
        Republic of Korea without a visa for a stay of up to 90 days if you are
        a tourist or on business.
      - >-
        Structural insulated panels (SIPs) are a high performance building
        system for residential and commercial construction. The panels consist
        of an insulating foam core sandwiched between two structural facings,
        typically oriented strand board (OSB). SIPs are manufactured under
        factory controlled conditions.
  - source_sentence: where are bussola shoes made?
    sentences:
      - >-
        According to Harvard University, biking at a moderate speed of 12 to
        13.9 miles per hour will cause a 155-pound person to burn 298 calories
        in 30 minutes. At a faster rate of 14 to 15.9 miles per hour, a person
        of the same weight will burn 372 calories.
      - >-
        If you had bought just one share of Microsoft at the IPO, you would now
        have 288 shares after all the splits. Those shares would be worth
        $44,505 at the current stock quote of $154.53. A $5,000 investment would
        have purchased 238 shares at the IPO price.
      - >-
        FRAM opens the first plant devoted exclusively to the development and
        manufacture of heavy duty air filters and cartridges, in Nevada,
        Missouri.
pipeline_tag: sentence-similarity
library_name: sentence-transformers

SentenceTransformer based on avsolatorio/GIST-small-Embedding-v0

This is a sentence-transformers model finetuned from avsolatorio/GIST-small-Embedding-v0. 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: avsolatorio/GIST-small-Embedding-v0
Maximum Sequence Length: 512 tokens
Output Dimensionality: 384 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': True}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, '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})
  (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("moshew/gist_small_ft_gooaq_v2")
# Run inference
sentences = [
    'where are bussola shoes made?',
    'FRAM opens the first plant devoted exclusively to the development and manufacture of heavy duty air filters and cartridges, in Nevada, Missouri.',
    'According to Harvard University, biking at a moderate speed of 12 to 13.9 miles per hour will cause a 155-pound person to burn 298 calories in 30 minutes. At a faster rate of 14 to 15.9 miles per hour, a person of the same weight will burn 372 calories.',
]
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]

Training Details

Training Dataset

Unnamed Dataset

Size: 2,000 training samples
Columns: sentence1, sentence2, and label
Approximate statistics based on the first 1000 samples:
sentence1 sentence2 label
type string string float
details
min: 8 tokens
mean: 12.05 tokens
max: 23 tokens

min: 13 tokens
mean: 59.28 tokens
max: 118 tokens

min: 0.0
mean: 0.5
max: 1.0

	sentence1	sentence2	label
type	string	string	float
details	min: 8 tokens mean: 12.05 tokens max: 23 tokens	min: 13 tokens mean: 59.28 tokens max: 118 tokens	min: 0.0 mean: 0.5 max: 1.0

Samples:

sentence1	sentence2	label
`what is the difference between rapid rise yeast and bread machine yeast?`	`Though there are some minor differences in shape and nutrients, Rapid-Rise Yeast is (pretty much) the same as Instant Yeast and Bread Machine Yeast. ... Also, Rapid-Rise Yeast is a little more potent than Active Dry Yeast and can be mixed in with your dry ingredients directly.`	`1.0`
`what is the difference between rapid rise yeast and bread machine yeast?`	`Application. To clarify, double-acting baking powder is “regular” baking powder. Single-acting baking powder exits, but when a recipe calls for baking powder it means double-acting. And even if a recipe does call for single-acting, you can substitute double-acting without worrying about it changing the recipe.`	`0.0`
`are light kits universal for ceiling fans?`	`Not all Universal Light Kits are actually Universal. They can be universal to only that manufacturer. ... Casablanca and Hunter Ceiling Fan Light Kits are universal only to their own fans.`	`1.0`

Loss: CoSENTLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 16
per_device_eval_batch_size: 16
num_train_epochs: 1
warmup_ratio: 0.1
seed: 12
bf16: True
dataloader_num_workers: 4

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: no
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: 1
eval_accumulation_steps: None
torch_empty_cache_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: 1
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: 12
data_seed: None
jit_mode_eval: False
use_ipex: False
bf16: True
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: 4
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}
tp_size: 0
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
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: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss
0.008	1	1.9382

Framework Versions

Python: 3.11.12
Sentence Transformers: 4.1.0
Transformers: 4.51.3
PyTorch: 2.6.0+cu124
Accelerate: 1.5.2
Datasets: 3.5.0
Tokenizers: 0.21.1

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},
}