metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:40906
  - loss:MatryoshkaLoss
  - loss:MegaBatchMarginLoss
widget:
  - source_sentence: >-
      One of three laminate structures that form the spindle pole body; the
      inner plaque is in the nucleus.
    sentences:
      - >-
        maturation of SSU-rRNA from tetracistronic rRNA transcript (SSU-rRNA,
        5.8S rRNA, 2S rRNA, LSU-rRNA)
      - leukotriene receptor activity
      - inner plaque of spindle pole body
  - source_sentence: >-
      The covalent attachment of a myristoyl group to the N-terminal amino acid
      residue of a protein.
    sentences:
      - MHC class I protein complex assembly
      - N-terminal protein myristoylation
      - neurotrophin receptor activity
  - source_sentence: >-
      The inner, i.e. lumen-facing, lipid bilayer of the plastid envelope; also
      faces the plastid stroma.
    sentences:
      - plastid inner membrane
      - neuron migration involved in retrograde extension
      - stomatal complex morphogenesis
  - source_sentence: >-
      Initiation of a region of tissue in a plant that is composed of one or
      more undifferentiated cells capable of undergoing mitosis and
      differentiation, thereby effecting growth and development of a plant by
      giving rise to more meristem or specialized tissue.
    sentences:
      - meristem initiation
      - polytene chromosome
      - cardiac ventricle development
  - source_sentence: >-
      The sex chromosome present in both sexes of species in which the male is
      the heterogametic sex. Two copies of the X chromosome are present in each
      somatic cell of females and one copy is present in males.
    sentences:
      - establishment of cell polarity involved in gastrulation cell migration
      - X chromosome
      - somatic diversification of immune receptors by N region addition
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - src2trg_accuracy
  - trg2src_accuracy
  - mean_accuracy
model-index:
  - name: SentenceTransformer
    results:
      - task:
          type: translation
          name: Translation
        dataset:
          name: Unknown
          type: unknown
        metrics:
          - type: src2trg_accuracy
            value: 0.7840546697038724
            name: Src2Trg Accuracy
          - type: trg2src_accuracy
            value: 0.7757023538344723
            name: Trg2Src Accuracy
          - type: mean_accuracy
            value: 0.7798785117691723
            name: Mean Accuracy
license: mit
language:
  - en
base_model:
  - Snowflake/snowflake-arctic-embed-m-v1.5
datasets:
  - NothingMuch/GO-Terms

SentenceTransformer

This is a sentence-transformers model trained on the parquet dataset. It maps sentences & paragraphs to a 128-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
Maximum Sequence Length: 512 tokens
Output Dimensionality: 128 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- parquet

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': 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})
  (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("NothingMuch/GO-Term-Embeddings")
# Run inference
sentences = [
    'The sex chromosome present in both sexes of species in which the male is the heterogametic sex. Two copies of the X chromosome are present in each somatic cell of females and one copy is present in males.',
    'X chromosome',
    'somatic diversification of immune receptors by N region addition',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 128]

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

Evaluation

Metrics

Translation

Evaluated with TranslationEvaluator

Metric	Value
src2trg_accuracy	0.7841
trg2src_accuracy	0.7757
mean_accuracy	0.7799

Training Details

Training Dataset

parquet

Dataset: parquet
Size: 40,906 training samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 8 tokens
mean: 42.05 tokens
max: 192 tokens

min: 3 tokens
mean: 10.48 tokens
max: 40 tokens

	anchor	positive
type	string	string
details	min: 8 tokens mean: 42.05 tokens max: 192 tokens	min: 3 tokens mean: 10.48 tokens max: 40 tokens

Samples:

anchor	positive
`Catalysis of the transfer of a mannose residue to an oligosaccharide, forming an alpha-(1->6) linkage.`	`1,6-alpha-mannosyltransferase activity`
`Catalysis of the hydrolysis of ester linkages within a single-stranded deoxyribonucleic acid molecule by creating internal breaks.`	`single-stranded DNA specific endodeoxyribonuclease activity`
`Catalysis of the hydrolysis of ester linkages within a single-stranded deoxyribonucleic acid molecule by creating internal breaks.`	`ssDNA-specific endodeoxyribonuclease activity`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MegaBatchMarginLoss",
    "matryoshka_dims": [
        64,
        32
    ],
    "matryoshka_weights": [
        1,
        1
    ],
    "n_dims_per_step": -1
}

Evaluation Dataset

parquet

Dataset: parquet
Size: 6,585 evaluation samples
Columns: anchor and positive
Approximate statistics based on the first 1000 samples:
anchor positive
type string string
details
min: 8 tokens
mean: 41.29 tokens
max: 253 tokens

min: 3 tokens
mean: 9.23 tokens
max: 44 tokens

	anchor	positive
type	string	string
details	min: 8 tokens mean: 41.29 tokens max: 253 tokens	min: 3 tokens mean: 9.23 tokens max: 44 tokens

Samples:

anchor	positive
`The maintenance of the structure and integrity of the mitochondrial genome; includes replication and segregation of the mitochondrial chromosome.`	`mitochondrial genome maintenance`
`The repair of single strand breaks in DNA. Repair of such breaks is mediated by the same enzyme systems as are used in base excision repair.`	`single strand break repair`
`Any process that modulates the frequency, rate or extent of DNA recombination, a DNA metabolic process in which a new genotype is formed by reassortment of genes resulting in gene combinations different from those that were present in the parents.`	`regulation of DNA recombination`

Loss: MatryoshkaLoss with these parameters:

{
    "loss": "MegaBatchMarginLoss",
    "matryoshka_dims": [
        64,
        32
    ],
    "matryoshka_weights": [
        1,
        1
    ],
    "n_dims_per_step": -1
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
per_device_train_batch_size: 32
per_device_eval_batch_size: 16
torch_empty_cache_steps: 250
learning_rate: 0.00025
lr_scheduler_type: cosine_with_restarts
warmup_steps: 25
seed: 25
load_best_model_at_end: True
batch_sampler: no_duplicates

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 32
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: 250
learning_rate: 0.00025
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: cosine_with_restarts
lr_scheduler_kwargs: {}
warmup_ratio: 0.0
warmup_steps: 25
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: 25
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: 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: 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
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
average_tokens_across_devices: False
prompts: None
batch_sampler: no_duplicates
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	Validation Loss	mean_accuracy
1.0016	641	0.2501	0.6276	0.7343
2.0016	1282	0.3146	0.5520	0.7651
2.9969	1920	0.1976	0.5097	0.7799

The bold row denotes the saved checkpoint.

Framework Versions

Python: 3.10.14
Sentence Transformers: 3.3.1
Transformers: 4.47.0
PyTorch: 2.4.0
Accelerate: 1.2.0
Datasets: 3.2.0
Tokenizers: 0.21.0

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",
}

MatryoshkaLoss

@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}
}

MegaBatchMarginLoss

@inproceedings{wieting-gimpel-2018-paranmt,
    title = "{P}ara{NMT}-50{M}: Pushing the Limits of Paraphrastic Sentence Embeddings with Millions of Machine Translations",
    author = "Wieting, John and Gimpel, Kevin",
    editor = "Gurevych, Iryna and Miyao, Yusuke",
    booktitle = "Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2018",
    address = "Melbourne, Australia",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/P18-1042",
    doi = "10.18653/v1/P18-1042",
    pages = "451--462",
}