SentenceTransformer based on Snowflake/snowflake-arctic-embed-l-v2.0

This is a sentence-transformers model finetuned from Snowflake/snowflake-arctic-embed-l-v2.0. 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-v2.0
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, 'architecture': 'XLMRobertaModel'})
  (1): Pooling({'word_embedding_dimension': 1024, '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("sentence_transformers_model_id")
# Run inference
sentences = [
    '[MASK] multiplied, in any case in which the service that was required has been performed in part, by the percentage which the length of service that was not performed is to the length of the service that was required to be performed. The amount will be determined under the following formula:',
    'the period of appointment begins, multiplied, in any case in which the service that was required has been performed in part, by the percentage which the length of service that was not performed is to the length of the service that was required to be performed. The amount will be determined under the following formula:',
    '[SUBSECTION c] Special enrollment periods.. A Part D eligible individual may enroll in a PDP or disenroll from a PDP and enroll in another PDP or MA-PD plan (as provided at § 422.62(b) of this chapter), as applicable, under any of the following circumstances: [CLAUSE 1] The individual involuntarily loses creditable prescription drug coverage or such coverage is involuntarily reduced so that it is no longer creditable coverage as defined under § 423.56(a). Loss of credible prescription drug coverage due to failure to pay any required premium is not considered involuntary loss of the coverage. [CLAUSE 2] The individual was not adequately informed, as required by standards established by CMS under § 423.56, that he or she has lost his or her creditable prescription drug coverage, that he or she never had credible prescription drug coverage, or the coverage is involuntarily reduced so that it is no longer creditable prescription drug coverage.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 1024]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[1.0000, 0.9334, 0.2975],
#         [0.9334, 1.0000, 0.2973],
#         [0.2975, 0.2973, 1.0000]])

Training Details

Training Dataset

Unnamed Dataset

Size: 24,880 training samples
Columns: sentence_0 and sentence_1
Approximate statistics based on the first 1000 samples:
sentence_0 sentence_1
type string string
details
min: 43 tokens
mean: 123.38 tokens
max: 340 tokens

min: 43 tokens
mean: 126.43 tokens
max: 366 tokens

	sentence_0	sentence_1
type	string	string
details	min: 43 tokens mean: 123.38 tokens max: 340 tokens	min: 43 tokens mean: 126.43 tokens max: 366 tokens

Samples:

sentence_0	sentence_1
`[SECTION HEADING] § [REF] Waiver of right of recovery. In determining whether there is good cause for waiver of any right of recovery which he may have against any applicant by reason of any payments made pursuant to a loan guarantee under this subpart, the Secretary shall take into consideration the extent to which:`	`[SECTION HEADING] § 57.1517 Waiver of right of recovery. In determining whether there is good cause for waiver of any right of recovery which he may have against any applicant by reason of any payments made pursuant to a loan guarantee under this subpart, the Secretary shall take into consideration the extent to which:`
`[SUBSECTION A] Any unmet deductible applied to the charges related to the reasonable costs that the facility incurs in providing the covered services; [CLAUSE 7] Rural health clinic services that meet the requirements set forth in part 491 of this chapter. [CITATIONS]`	`[SUBSECTION A] Any unmet deductible applied to the charges related to the reasonable costs that the facility incurs in providing the covered services; [CLAUSE 7] Rural health clinic services that meet the requirements set forth in part 491 of this chapter. [CITATIONS]`
[SUBSECTION b] Waiver of the right to appear.. (1) An enrollee may submit to OMHA a written statement indicating that he or she does not wish to appear at the hearing. [ITEM i] For expedited hearings, an enrollee may indicate in writing or orally [MASK] ii] The OMHA hearing office must document all oral waivers in writing and maintain the documentation in the case files. [CLAUSE 2] The enrollee may subsequently withdraw his or her waiver in writing at any time before the notice of the hearing decision is issued; however, by withdrawing the waiver the enrollee agrees to an extension of the adjudication period as specified in § 423.2016, that may be necessary to schedule and hold the hearing.	[SUBSECTION b] Waiver of the right to appear.. (1) An enrollee may submit to OMHA a written statement indicating that he or she does not wish to appear at the hearing. [ITEM i] For expedited hearings, an enrollee may indicate in writing or orally that he or she does not wish to appear at the hearing. [ITEM ii] The OMHA hearing office must document all oral waivers in writing and maintain the documentation in the case files. [CLAUSE 2] The enrollee may subsequently withdraw his or her waiver in writing at any time before the notice of the hearing decision is issued; however, by withdrawing the waiver the enrollee agrees to an extension of the adjudication period as specified in § 423.2016, that may be necessary to schedule and hold the hearing.

Loss: DenoisingAutoEncoderLoss

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 16
per_device_eval_batch_size: 16
multi_dataset_batch_sampler: round_robin

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
num_train_epochs: 3
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}
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}
parallelism_config: 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: None
hub_always_push: False
hub_revision: None
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
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: False
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: round_robin
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss
0.3215	500	6.6865
0.6431	1000	5.7788
0.9646	1500	5.5835
1.2862	2000	5.3876
1.6077	2500	5.2766
1.9293	3000	5.2108
2.2508	3500	5.1317
2.5723	4000	5.0701
2.8939	4500	5.0288

Framework Versions

Python: 3.12.6
Sentence Transformers: 5.2.0
Transformers: 4.56.0
PyTorch: 2.8.0+cu129
Accelerate: 1.10.1
Datasets: 4.4.1
Tokenizers: 0.22.0

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",
}

DenoisingAutoEncoderLoss

@inproceedings{wang-2021-TSDAE,
    title = "TSDAE: Using Transformer-based Sequential Denoising Auto-Encoderfor Unsupervised Sentence Embedding Learning",
    author = "Wang, Kexin and Reimers, Nils and Gurevych, Iryna",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2021",
    month = nov,
    year = "2021",
    address = "Punta Cana, Dominican Republic",
    publisher = "Association for Computational Linguistics",
    pages = "671--688",
    url = "https://arxiv.org/abs/2104.06979",
}

Downloads last month: 3

Safetensors

Model size

0.6B params

Tensor type

F32

Model tree for atx-labs/snowflake-custom-noise-tsdae-cfr-finetuned

Base model

Snowflake/snowflake-arctic-embed-l-v2.0

Finetuned

(19)

this model