This is a sentence-transformers model finetuned from sentence-transformers/all-mpnet-base-v2. It maps sentences & paragraphs to a 768-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-mpnet-base-v2
Maximum Sequence Length: 384 tokens
Output Dimensionality: 768 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': 384, 'do_lower_case': False}) with Transformer model: MPNetModel 
  (1): Pooling({'word_embedding_dimension': 768, '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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("CatkinChen/all-mpnet-base-v2_retrieval_finetuned_2025-03-29_23-37-00")
# Run inference
sentences = [
    'Which Hogwarts house does Harry get sorted into?',
    'Book: 1, Chapter: 7\nPassage: "Are you sure? You could be great, you know, its all here in your head, and Slytherin will help you on the way to greatness, no doubt about that - no? Well, if you\'re sure - better be GRYFFINDOR!" Harry heard the hat shout the last word to the whole hall. He took off the hat and walked shakily toward the Gryffindor table. He was so relieved to have been chosen and not put in Slytherin, he hardly noticed that he was getting the loudest cheer yet.',
    'Book: 4, Chapter: 9\nPassage: "So," said Mr. Diggory with a kind of savage triumph, looking down upon Winky, who was still shaking convulsively. "I is not doing it!" she squealed, her eyes rolling in terror. "I is not, I is not, I is not knowing how! I is a good elf, I isn\'t using wands, I isn\'t knowing how!" "You\'ve been caught red-handed, elf!" Mr. Diggory roared.',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

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

Evaluation

Metrics

Triplet

Dataset: test-eval
Evaluated with TripletEvaluator

Metric	Value
cosine_accuracy	0.8286

Training Details

Training Dataset

Unnamed Dataset

Size: 309 training samples
Columns: sentence_0, sentence_1, and sentence_2
Approximate statistics based on the first 309 samples:
sentence_0 sentence_1 sentence_2
type string string string
details
min: 9 tokens
mean: 17.74 tokens
max: 53 tokens
min: 46 tokens
mean: 206.97 tokens
max: 384 tokens
min: 44 tokens
mean: 180.12 tokens
max: 384 tokens

Samples:

sentence_0	sentence_1	sentence_2
`What is the name of the spell that causes a person to be killed?`	Book: 4, Chapter: 14[object Object]Passage: Hermione's hand shook slightly as, for the third time, she raised it into the air. "Yes?" said Moody, looking at her. "Avada Kedavra," Hermione whispered. Several people looked uneasily around at her, including Ron. "Ah," said Moody, another slight smile twisting his lopsided mouth. "Yes, the last and worst. Avada Kedavra ... the Killing Curse." He put his hand into the glass jar, and almost as though it knew what was coming, the third spider scuttled frantically around the bottom of the jar, trying to evade Moody's fingers, but he trapped it, and placed it upon the desktop. It started to scuttle frantically across the wooden surface. Moody raised his wand, and Harry felt a sudden thrill of foreboding. "Avada Kedavra!" Moody roared. There was a flash of blinding green light and a rushing sound, as though a vast, invisible something was soaring through the air - instantaneously the spider rolled over onto its back, unmarked, but unmistakably dead. Several o...	`Book: 6, Chapter: 24[object Object]Passage: A roar of celebration erupted from the hole behind her. Harry gaped as people began to scream at the sight of him; several hands pulled him into the room. "We won!" yelled Ron, bounding into sight and brandishing the silver Cup at Harry. "We won! Four hundred and fifty to a hundred and forty! We won!" Harry looked around; there was Ginny running toward him; she had a hard, blazing look in her face as she threw her arms around him.`
`Who is the first person Harry meets from the wizarding world?`	`Book: 1, Chapter: 4[object Object]Passage: Harry looked up at the giant. He meant to say thank you, but the words got lost on the way to his mouth, and what he said instead was, "Who are you?" The giant chuckled. "True, I haven't introduced meself. Rubeus Hagrid, Keeper of Keys and Grounds at Hogwarts." He held out an enormous hand and shook Harry's whole arm. "What about that tea then, eh?" he said, rubbing his hands together. "I'd not say no ter summat stronger if yeh've got it, mind."`	Book: 2, Chapter: 17[object Object]Passage: He was standing at the end of a very long, dimly lit chamber. Towering stone pillars entwined with more carved serpents rose to support a ceiling lost in darkness, casting long, black shadows through the odd, greenish gloom that filled the place. His heart beating very fast, Harry stood listening to the chill silence. Could the basilisk be lurking in a shadowy corner, behind a pillar? And where was Ginny? He pulled out his wand and moved forward between the serpentine columns. Every careful footstep echoed loudly off the shadowy walls. He kept his eyes narrowed, ready to clamp them shut at the smallest sign of movement. The hollow eye sockets of the stone snakes seemed to be following him. More than once, with a jolt of the stomach, he thought he saw one stir. Then, as he drew level with the last pair of pillars, a statue high as the Chamber itself loomed into view, standing against the back wall. Harry had to crane his neck to look up into the giant face ...
`Who is the Half-Blood Prince?`	Book: 6, Chapter: 28[object Object]Passage: Harry uttered an inarticulate yell of rage: In that instant, he cared not whether he lived or died. Pushing himself to his feet again, he staggered blindly toward Snape, the man he now hated as much as he hated Voldemort himself -[object Object]"Sectum - !" Snape flicked his wand and the curse was repelled yet again; but Harry was mere feet away now and he could see Snape's face clearly at last: He was no longer sneering or jeering; the blazing flames showed a face full of rage. Mustering all his powers of concentration, Harry thought, Levi -[object Object]"No, Potter!" screamed Snape. There was a loud BANG and Harry was soaring backward, hitting the ground hard again, and this time his wand flew out of his hand. He could hear Hagrid yelling and Fang howling as Snape closed in and looked down on him where he lay, wandless and defenseless as Dumbledore had been. Snape's pale face, illuminated by the flaming cabin, was suffused with hatred just as it had been before he had cursed Dumbl...	Book: 3, Chapter: 17[object Object]Passage: Lupin forced a laugh. "You're the cleverest witch of your age I've ever met, Hermione." "I'm not," Hermione whispered. "If I'd been a bit cleverer, I'd have told everyone what you are!" "But they already know," said Lupin. "At least, the staff do." "Dumbledore hired you when he knew you were a werewolf?" Ron gasped. "Is he mad?" "Some of the staff thought so," said Lupin. "He had to work very hard to convince certain teachers that I'm trustworthy -"[object Object]"AND HE WAS WRONG!"

Loss: TripletLoss with these parameters:
json
1{ 2 "distance_metric": "TripletDistanceMetric.COSINE", 3 "triplet_margin": 0.5 4}

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 4
per_device_eval_batch_size: 4
num_train_epochs: 1
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: 4
per_device_eval_batch_size: 4
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: 1
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
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: round_robin

Training Logs

Epoch	Step	test-eval_cosine_accuracy
0.1282	10	0.8000
0.2564	20	0.8000
0.3846	30	0.8071
0.5128	40	0.8214
0.6410	50	0.8286
0.7692	60	0.8286
0.8974	70	0.8286
1.0	78	0.8286
-1	-1	0.8286

Framework Versions

Python: 3.12.2
Sentence Transformers: 3.4.1
Transformers: 4.50.0
PyTorch: 2.4.1
Accelerate: 1.4.0
Datasets: 2.19.2
Tokenizers: 0.21.0

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",
}

TripletLoss

bibtex

@misc{hermans2017defense,
    title={In Defense of the Triplet Loss for Person Re-Identification},
    author={Alexander Hermans and Lucas Beyer and Bastian Leibe},
    year={2017},
    eprint={1703.07737},
    archivePrefix={arXiv},
    primaryClass={cs.CV}
}

	sentence_0	sentence_1	sentence_2
type	string	string	string
details	min: 9 tokens mean: 17.74 tokens max: 53 tokens	min: 46 tokens mean: 206.97 tokens max: 384 tokens	min: 44 tokens mean: 180.12 tokens max: 384 tokens