This is a sentence-transformers model finetuned from BAAI/bge-base-en-v1.5. 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: BAAI/bge-base-en-v1.5
Maximum Sequence Length: 512 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': 512, 'do_lower_case': True}) 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.

python

from sentence_transformers import SentenceTransformer

# Download from the 🤗 Hub
model = SentenceTransformer("CatkinChen/BAAI_bge-base-en-v1.5_retrieval_finetuned_2025-03-29_23-58-08")
# Run inference
sentences = [
    'What is the name of the magical creature that sings underwater?',
    'Book: 6, Chapter: 30\nPassage: He wondered whether Scrimgeour or any of these important people were really sorry that Dumbledore was dead. But then he heard music, strange, otherworldly music, and he forgot his dislike of the Ministry in looking around for the source of it. He was not the only one: Many heads were turning, searching, a little alarmed. "In there," whispered Ginny in Harry\'s ear. And he saw them in the clear green sunlit water, inches below the surface, reminding him horribly of the Inferi: a chorus of merpeople singing in a strange language he did not understand, their pallid faces rippling, their purplish hair flowing all around them. The music made the hair on Harry\'s neck stand up, and yet it was not unpleasant.',
    'Book: 6, Chapter: 29\nPassage: We can arrange transport home afterward." "Seconded," barked Professor Sprout. "I suppose ... yes ..." said Slughorn in a rather agitated voice, while Hagrid let out a strangled sob of assent. "He\'s coming," said Professor McGonagall suddenly, gazing down into the grounds. "The Minister ... and by the looks of it, he\'s brought a delegation ..."\n"Can I leave, Professor?" said Harry at once. He had no desire at all to see, or be interrogated by, Rufus Scrimgeour tonight. "You may," said Professor McGonagall. "And quickly."',
]
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	0.8214

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: 212.92 tokens
max: 512 tokens
min: 41 tokens
mean: 175.74 tokens
max: 512 tokens

Samples:

sentence_0	sentence_1	sentence_2
`What is the name of the charm that creates a shield?`	Book: 7, Chapter: 36[object Object]Passage: screamed Mrs. Weasley. Bellatrix laughed, the same exhilarated laugh her cousin Sirius had given as he toppled backward through the veil, and suddenly Harry knew what was going to happen before it did. Molly's curse soared beneath Bellatrix's outstretched arm and hit her squarely in the chest, directly over her heart. Bellatrix's gloating smile froze, her eyes seemed to bulge: For the tiniest space of time she knew what had happened, and then she toppled, and the watching crowd roared, and Voldemort screamed. Harry felt as though he turned in slow motion; he saw McGonagall, Kingsley, and Slughorn blasted backward, flailing and writhing through the air, as Voldemort's fury at the fall of his last, best lieutenant exploded with the force of a bomb. Voldemort raised his wand and directed it at Molly Weasley. "Protego!" roared Harry, and the Shield Charm expanded in the middle of the Hall, and Voldemort stared around for the source as Harry pulled off the Invi...	Book: 4, Chapter: 24[object Object]Passage: "Can I have one of these cakes, Hagrid?" "Help yerself," said Hagrid, wiping his eyes on the back of his hand. "Ar, he's righ', o' course - yeh're all righ' ... I bin stupid ... my ol' dad woulda bin ashamed o' the way I've bin behavin'. ..." More tears leaked out, but he wiped them away more forcefully, and said, "Never shown you a picture of my old dad, have I? Here ..."[object Object]Hagrid got up, went over to his dresser, opened a drawer, and pulled out a picture of a short wizard with Hagrid's crinkled black eyes, beaming as he sat on top of Hagrid's shoulder. Hagrid was a good seven or eight feet tall, judging by the apple tree beside him, but his face was beardless, young, round, and smooth - he looked hardly older than eleven. "Tha' was taken jus' after I got inter Hogwarts," Hagrid croaked.
`Who is the author of 'Advanced Potion-Making'?`	Book: 6, Chapter: 9[object Object]Passage: You can use ingredients from the store cupboard today, and I'm sure we can lend you some scales, and we've got a small stock of old books here, they'll do until you can write to Flourish and Blotts. ..."[object Object]Slughorn strode over to a corner cupboard and, after a moment's foraging, emerged with two very battered-looking copies of Advanced Potion-Making by Libatius Borage, which he gave to Harry and Ron along with two sets of tarnished scales. "Now then," said Slughorn, returning to the front of the class and inflating his already bulging chest so that the buttons on his waistcoat threatened to burst off, "I've prepared a few potions for you to have a look at, just out of interest, you know. These are the kind of thing you ought to be able to make after completing your N.E.W.T.s. You ought to have heard of 'em, even if you haven't made 'em yet. Anyone tell me what this one is?" He indicated the cauldron nearest the Slytherin table. Harry raised himself slightly i...	Book: 7, Chapter: 15[object Object]Passage: Xeno is printing all the stuff the Prophet's ignoring, not a single mention of Crumple-Horned Snorkacks in the last issue. How long they'll let him get away with it, mind, I don't know. But Xeno says, front page of every issue, that any wizard who's against You-Know-Who ought to make helping Harry Potter their number-one priority." "Hard to help a boy who's vanished off the face of the earth," said Dirk. "Listen, the fact that they haven't caught him yet's one hell of an achievement," said Ted. "I'd take tips from him gladly; it's what we're trying to do, stay free, isn't it?" "Yeah, well, you've got a point there," said Dirk heavily.
`What is the name of the spell that causes a person to shrink?`	Book: 4, Chapter: 14[object Object]Passage: No sound came from it, but Harry was sure that if it could have given voice, it would have been screaming. Moody did not remove his wand, and the spider started to shudder and jerk more violently -[object Object]"Stop it!" Hermione said shrilly. Harry looked around at her. She was looking, not at the spider, but at Neville, and Harry, following her gaze, saw that Neville's hands were clenched upon the desk in front of him, his knuckles white, his eyes wide and horrified. Moody raised his wand. The spider's legs relaxed, but it continued to twitch. "Reducio," Moody muttered, and the spider shrank back to its proper size. He put it back into the jar.	`Book: 7, Chapter: 30[object Object]Passage: he shrieked, standing over his sister and smacking himself on the forehead with his fist. "We haven't got him, and they've gorn and killed her!" "She's only Stunned," said Professor McGonagall impatiently, who had stooped down to examine Alecto. "She'll be perfectly all right." "No she bludgering well won't!" bellowed Amycus. "Not after the Dark Lord gets hold of her!`

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: 16
per_device_eval_batch_size: 16
num_train_epochs: 10
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: 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: 10
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
0.5	10	0.8071
1.0	20	0.8286
1.5	30	0.8357
2.0	40	0.8429
2.5	50	0.8357
3.0	60	0.8643
3.5	70	0.8500
4.0	80	0.8429
4.5	90	0.8643
5.0	100	0.8500
5.5	110	0.8286
6.0	120	0.8214
6.5	130	0.8143
7.0	140	0.8214
7.5	150	0.8143
8.0	160	0.8214
8.5	170	0.8214
9.0	180	0.8214
9.5	190	0.8214
10.0	200	0.8214
-1	-1	0.8214

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: 212.92 tokens max: 512 tokens	min: 41 tokens mean: 175.74 tokens max: 512 tokens