This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-l. It maps sentences & paragraphs to a 1024-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: Snowflake/snowflake-arctic-embed-l
Maximum Sequence Length: 512 tokens
Output Dimensionality: 1024 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': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 1024, '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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("JTh34/puppy-embed-8985966a")
# Run inference
sentences = [
    'How can I ensure my puppy feels safe around my young child to prevent defensive biting?',
    'Until the ages of 5 to 7, children have trouble grasping others’ feelings — whether the others are people or pets. Though a tight squeeze may signal love from your 3-year-old to you, it can instill panic or frighten a young pup. A squeezed puppy may feel trapped and bite defensively, even if under normal circumstances they wouldn’t react this way. This puppy can mature into a dog with an innate fear of young children or into a dog who’s immediately tense in their company.\n\nIf you have a needy toddler, postpone getting a puppy for at least a couple of years. Your child needs all your attention to develop a strong sense of self. A puppy will pull you away from your parental duties and rival the toddler for your attention. This situation is almost guaranteed to cause some headaches. Introduce the puppy after your child is more emotionally steady and is also excited and ready for the addition.\n\nIf you’re convinced this is the right time to add a puppy to your family, consider the option of an older puppy who has been well socialized to children during their first few months. Find a 6- to 10-month-old puppy who has been given up for reasons such as a move or human allergies to the puppy. After they’re past the intense nipping phase, puppies are less likely to think of the kids as littermates, also known as siblings. As a bonus, an older puppy has better bladder control and may already be completely house-trained.\n\nAre you ready to take the plunge? Flip to Chapter 8 for tips on raising kids and puppies together.',
    'To avoid arguments, discuss ahead of time how you’ll raise this puppy and which rules make the most sense for your lifestyle and future situation. Have a heartfelt talk with your partner — ideally, before bringing your puppy home. Be sure to discuss the following topics:\n\n» What are your separate visions and hopes for adding a dog to your life? » Where will the puppy sleep? » How will you share responsibilities, from feeding and walking to exercise and\n\n» What are your feeding philosophies, from kibbles in the bowl to handouts\n\n» How much money will you apportion to training, grooming, and maintaining\n\nYour puppy will live and love most serenely in a household where you can both agree.\n\nIf you and your partner are retired empty nesters, now may be a terrific time to get a puppy, because you’re probably home more often and have time to be attentive to your puppy’s schedule. However, remember that a young puppy’s needs can be quite demanding. If you already completed the 4 a.m. diaper-changing routine and you’d rather skip these experiences than relive them, consider an older puppy who has been in a supportive environment or one whom the breeder is willing to socialize and train for you. Also, pay close attention to the exercise requirements when choosing a puppy. You need to ensure that you and your soon-to-be-adult dog are a solid match.\n\nIf you’re getting a puppy to raise with your young child, you may suddenly feel like you have twins (except that one is slightly furrier than the other). Raising a puppy with children younger than the age of 5 is a tremendous undertaking, and it’s one that often creates more stress than it’s worth.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities.shape)
# [3, 3]

Evaluation

Metrics

Information Retrieval

Evaluated with InformationRetrievalEvaluator

Metric	Value
cosine_accuracy@1	0.7273
cosine_accuracy@3	0.8636
cosine_accuracy@5	0.9545
cosine_accuracy@10	1.0
cosine_precision@1	0.7273
cosine_precision@3	0.2879
cosine_precision@5	0.1909
cosine_precision@10	0.1
cosine_recall@1	0.7273
cosine_recall@3	0.8636
cosine_recall@5	0.9545
cosine_recall@10	1.0
cosine_ndcg@10	0.8667
cosine_mrr@10	0.8235
cosine_map@100	0.8235

Training Details

Training Dataset

Unnamed Dataset

Size: 105 training samples
Columns: sentence_0 and sentence_1
Approximate statistics based on the first 105 samples:
sentence_0 sentence_1
type string string
details
min: 13 tokens
mean: 21.33 tokens
max: 32 tokens
min: 225 tokens
mean: 340.84 tokens
max: 512 tokens

Samples:

sentence_0	sentence_1
`What are the best practices for training my puppy effectively?`	Copyright © 2019 by John Wiley & Sons, Inc., Hoboken, New Jersey[object Object][object Object]No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted under Sections 107 or 108 of the 1976 United States Copyright Act, without the prior written permission of the Publisher. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at [object Object].[object Object][object Object]Trademarks: Wiley, For Dummies, the Dummies Man logo, Dummies.com, Making Everything Easier, and related trade dress are trademarks or registered trademarks of John Wiley & Sons, Inc. and may not be used without written permission. All other trademarks are the property of their respective owners. John Wiley & Sons, Inc. is not associated with any product or...
`How can I ensure that I am following copyright guidelines when using puppy care resources?`	Copyright © 2019 by John Wiley & Sons, Inc., Hoboken, New Jersey[object Object][object Object]No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning or otherwise, except as permitted under Sections 107 or 108 of the 1976 United States Copyright Act, without the prior written permission of the Publisher. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748-6011, fax (201) 748-6008, or online at [object Object].[object Object][object Object]Trademarks: Wiley, For Dummies, the Dummies Man logo, Dummies.com, Making Everything Easier, and related trade dress are trademarks or registered trademarks of John Wiley & Sons, Inc. and may not be used without written permission. All other trademarks are the property of their respective owners. John Wiley & Sons, Inc. is not associated with any product or...
`What should I do if I need professional assistance with my puppy's training or care?`	LIMIT OF LIABILITY/DISCLAIMER OF WARRANTY: THE PUBLISHER AND THE AUTHOR MAKE NO REPRESENTATIONS OR WARRANTIES WITH RESPECT TO THE ACCURACY OR COMPLETENESS OF THE CONTENTS OF THIS WORK AND SPECIFICALLY DISCLAIM ALL WARRANTIES, INCLUDING WITHOUT LIMITATION WARRANTIES OF FITNESS FOR A PARTICULAR PURPOSE. NO WARRANTY MAY BE CREATED OR EXTENDED BY SALES OR PROMOTIONAL MATERIALS. THE ADVICE AND STRATEGIES CONTAINED HEREIN MAY NOT BE SUITABLE FOR EVERY SITUATION. THIS WORK IS SOLD WITH THE UNDERSTANDING THAT THE PUBLISHER IS NOT ENGAGED IN RENDERING LEGAL, ACCOUNTING, OR OTHER PROFESSIONAL SERVICES. IF PROFESSIONAL ASSISTANCE IS REQUIRED, THE SERVICES OF A COMPETENT PROFESSIONAL PERSON SHOULD BE SOUGHT. NEITHER THE PUBLISHER NOR THE AUTHOR SHALL BE LIABLE FOR DAMAGES ARISING HEREFROM. THE FACT THAT AN ORGANIZATION OR WEBSITE IS REFERRED TO IN THIS WORK AS A CITATION AND/OR A POTENTIAL SOURCE OF FURTHER INFORMATION DOES NOT MEAN THAT THE AUTHOR OR THE PUBLISHER ENDORSES THE INFORMATION THE ORG...

Loss: MatryoshkaLoss with these parameters:
json
1{ 2 "loss": "MultipleNegativesRankingLoss", 3 "matryoshka_dims": [ 4 1024, 5 512, 6 256, 7 128, 8 64 9 ], 10 "matryoshka_weights": [ 11 1, 12 1, 13 1, 14 1, 15 1 16 ], 17 "n_dims_per_step": -1 18}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 2
per_device_eval_batch_size: 2
num_train_epochs: 5
multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 2
per_device_eval_batch_size: 2
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
num_train_epochs: 5
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: {}
warmup_ratio: 0.0
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}
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: round_robin

Training Logs

Epoch	Step	cosine_ndcg@10
0.9434	50	0.7752
1.0	53	0.7752
1.8868	100	0.7808
2.0	106	0.8193
2.8302	150	0.8439
3.0	159	0.8439
3.7736	200	0.8607
4.0	212	0.8667
4.7170	250	0.8667
5.0	265	0.8667

Framework Versions

Python: 3.12.10
Sentence Transformers: 4.1.0
Transformers: 4.51.3
PyTorch: 2.7.0
Accelerate: 1.6.0
Datasets: 3.6.0
Tokenizers: 0.21.1

Citation

BibTeX

Sentence Transformers

bibtex

@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",
}

MatryoshkaLoss

bibtex

@misc{kusupati2024matryoshka,
    title={Matryoshka Representation Learning},
    author={Aditya Kusupati and Gantavya Bhatt and Aniket Rege and Matthew Wallingford and Aditya Sinha and Vivek Ramanujan and William Howard-Snyder and Kaifeng Chen and Sham Kakade and Prateek Jain and Ali Farhadi},
    year={2024},
    eprint={2205.13147},
    archivePrefix={arXiv},
    primaryClass={cs.LG}
}

MultipleNegativesRankingLoss

bibtex

@misc{henderson2017efficient,
    title={Efficient Natural Language Response Suggestion for Smart Reply},
    author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
    year={2017},
    eprint={1705.00652},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}

	sentence_0	sentence_1
type	string	string
details	min: 13 tokens mean: 21.33 tokens max: 32 tokens	min: 225 tokens mean: 340.84 tokens max: 512 tokens