SentenceTransformer based on sentence-transformers/all-MiniLM-L6-v2

This is a sentence-transformers model finetuned from sentence-transformers/all-MiniLM-L6-v2 on the train dataset. 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: sentence-transformers/all-MiniLM-L6-v2
Maximum Sequence Length: 256 tokens
Output Dimensionality: 384 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- train

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': 256, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 384, '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})
  (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("TONKKrongyuth/finetune-all-minilm-L6-v2-proofwiki_wo-theorem")
# Run inference
sentences = [
    ':$\\displaystyle \\sum_{n \\mathop = 1}^\\infty \\dfrac 1 {x_n \\map {J_0 } {x_n} } = 0 \\cdotp 38479 \\ldots$where::$x_n$ is the $n$th [[Definition:Zero of Function|zero]] of the [[Definition:Bessel Function of the First Kind|order $1$ Bessel function of the first kind]]:$\\map {J_0 } {x_n}$ is the [[Definition:Bessel Function of the First Kind|order $0$ Bessel function of the first kind]] of $x_n$.',
    "A '''Bessel function of the first kind of order $n$''' is a [[Definition:Bessel Function|Bessel function]] which is [[Definition:Non-Singular Point|non-singular]] at the [[Definition:Origin|origin]]It is usually denoted $\\map {J_n} x$, where $x$ is the [[Definition:Dependent Variable|dependent variable]] of the instance of '''[[Definition:Bessel's Equation|Bessel's equation]]''' to which $\\map {J_n} x$ forms a [[Definition:Solution of Differential Equation|solution]].",
    "=== [[Definition:Prime Number/Definition 1|Definition 1]] ==={{:Definition:Prime Number/Definition 1}}=== [[Definition:Prime Number/Definition 2|Definition 2]] ==={{:Definition:Prime Number/Definition 2}}=== [[Definition:Prime Number/Definition 3|Definition 3]] ==={{:Definition:Prime Number/Definition 3}}=== [[Definition:Prime Number/Definition 4|Definition 4]] ==={{:Definition:Prime Number/Definition 4}}=== [[Definition:Prime Number/Definition 5|Definition 5]] ==={{:Definition:Prime Number/Definition 5}}=== [[Definition:Prime Number/Definition 6|Definition 6]] ==={{:Definition:Prime Number/Definition 6}}=== [[Definition:Prime Number/Definition 7|Definition 7]] ==={{:Definition:Prime Number/Definition 7}}=== Euclid's Definition ==={{EuclidSaid}}:''{{:Definition:Euclid's Definitions - Book VII/11 - Prime Number}}''{{EuclidDefRefNocat|VII|11|Prime Number}}",
]
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

train

Dataset: train
Size: 20,935 training samples
Columns: theorems_content, refs_content, and score

Approximate statistics based on the first 1000 samples:

	theorems_content	refs_content	score
type	string	string	float
details	min: 16 tokens mean: 129.25 tokens max: 256 tokens	min: 12 tokens mean: 141.91 tokens max: 256 tokens	min: 0.0 mean: 0.49 max: 1.0

Samples:

theorems_content	refs_content	score
`Let $C_n$ be the [[Definition:Cyclic Group`	cyclic group]] of [[Definition:Order of Structure	order]] $n$.Let $C_n = \gen a$, that is, that $C_n$ is [[Definition:Generator of Cyclic Group
`Let $T = \struct {S, \tau}$ be a [[Definition:Topological Space`	topological space]] where $\tau$ is the [[Definition:Discrete Topology	discrete topology]] on $S$.Then $T$ is a [[Definition:Scattered Space
`Let $A \subseteq \R$ be the [[Definition:Set`	set]] of all points on $\R$ defined as::$A := \set 0 \cup \set {\dfrac 1 n : n \in \Z_{>0} }$Let $\struct {A, \tau_d}$ be the [[Definition:Integer Reciprocal Space	integer reciprocal space]] with [[Definition:Zero (Number)

Loss: CosineSimilarityLoss with these parameters:

{
    "loss_fct": "torch.nn.modules.loss.MSELoss"
}

Evaluation Dataset

train

Dataset: train
Size: 5,193 evaluation samples
Columns: theorems_content, refs_content, and score

Approximate statistics based on the first 1000 samples:

	theorems_content	refs_content	score
type	string	string	float
details	min: 22 tokens mean: 134.45 tokens max: 256 tokens	min: 11 tokens mean: 139.82 tokens max: 256 tokens	min: 0.0 mean: 0.5 max: 1.0

Samples:

theorems_content	refs_content	score
`Let $F$ be a [[Definition:Field (Abstract Algebra)`	field]].Then $F$ is a [[Definition:Principal Ideal Domain	principal ideal domain]].
`:$G$ is a [[Definition:Left Ideal`	left ideal]] of $\struct {\map {\MM_S} 2, +, \times}$.	`$J$ is a '''left ideal of $R$''' {{iff}}::$\forall j \in J: \forall r \in R: r \circ j \in J$that is, {{iff}}::$\forall r \in R: r \circ J \subseteq J$`
`Let $S$ be a [[Definition:Set`	set]].Let $T \subseteq S$ be a given [[Definition:Subset	subset]] of $S$.Let $\powerset S$ denote the [[Definition:Power Set

Loss: CosineSimilarityLoss with these parameters:

{
    "loss_fct": "torch.nn.modules.loss.MSELoss"
}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: epoch
push_to_hub: True
hub_model_id: TONKKrongyuth/finetune-all-minilm-L6-v2-proofwiki_wo-theorem

All Hyperparameters

Click to expand

overwrite_output_dir: False
do_predict: False
eval_strategy: epoch
prediction_loss_only: True
per_device_train_batch_size: 8
per_device_eval_batch_size: 8
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: 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}
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: True
resume_from_checkpoint: None
hub_model_id: TONKKrongyuth/finetune-all-minilm-L6-v2-proofwiki_wo-theorem
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: batch_sampler
multi_dataset_batch_sampler: proportional

Training Logs

Epoch	Step	Training Loss	train loss
1.0	2617	0.0357	0.0275
2.0	5234	0.0171	0.0247
3.0	7851	0.0097	0.0237

Framework Versions

Python: 3.11.11
Sentence Transformers: 3.4.1
Transformers: 4.48.3
PyTorch: 2.6.0+cu124
Accelerate: 1.3.0
Datasets: 3.3.2
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",
}

Downloads last month: 2

Safetensors

Model size

22.7M params

Tensor type

F32

Model tree for TONKKrongyuth/finetune-all-minilm-L6-v2-proofwiki_wo-theorem

Base model

sentence-transformers/all-MiniLM-L6-v2

Finetuned

(822)

this model

Paper for TONKKrongyuth/finetune-all-minilm-L6-v2-proofwiki_wo-theorem

Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks

Paper • 1908.10084 • Published Aug 27, 2019 • 12