SentenceTransformer based on sentence-transformers/all-mpnet-base-v2

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 retrieval.

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
Supported Modality: Text
Training Dataset: Poem_classification - train_data.csv & Poem_classification - test_data.csv from kaggle
Loss : BatchAllTripletLoss
Fine tuning type : supervised metric learning

Model Sources

Documentation: Sentence Transformers Documentation
Repository: Sentence Transformers on GitHub
Hugging Face: Sentence Transformers on Hugging Face

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'MPNetModel'})
  (1): Pooling({'embedding_dimension': 768, 'pooling_mode': 'mean', '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("auukjkjk/poem_embedding_model")
# Run inference
sentences = [
    'Six months ago, the measuring of whiskey left in the jug, urine on the mattress, couch cushions, the crotch of pants in wear. You watch how breath lifts a chest, how a person breathes— sick hobbies of when we must. You watch how you become illiterate at counting. Six or seven broken breathalyzers; a joke formulates in',
    'Music',
    'Affection',
]
embeddings = model.encode(sentences)
print(embeddings.shape)
# [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(embeddings, embeddings)
print(similarities)
# tensor([[ 1.0000,  0.2384, -0.0032],
#         [ 0.2384,  1.0000,  0.1950],
#         [-0.0032,  0.1950,  1.0000]])

Training Details

Training Dataset

Unnamed Dataset

Size: 991 training samples
Columns: sentence and label
Approximate statistics based on the first 100 samples:
sentence label
type string int

modality text
details
min: 11 tokens
mean: 65.32 tokens
max: 97 tokens

0: ~27.88%
1: ~26.92%
2: ~25.00%
3: ~20.19%

	sentence	label
type	string	int
modality	text
details	min: 11 tokens mean: 65.32 tokens max: 97 tokens	0: ~27.88% 1: ~26.92% 2: ~25.00% 3: ~20.19%

Samples:

sentence	label
`What is it you feel I asked Kurt when you listen toRavel’s String Quartet in F-major, his face was so lit upand I wondered, “the music is unlike the world I liveor think in, it’s from somewhere else, unfamiliar and unknown,not because it is relevant to the familiar and comfortable,but`	`3`
`I have been a spendthrift Dropping from lazy fingers Quiet coloured hours, Fluttering away from me Like oak and beech leaves in October.I have lived keenly and wastefully, Like a bush or a sun insect— Lived sensually and thoughtfully, Loving the flesh and the beauty of this world— Green ivy about ruined towers, The out-pouring of the grey sea, And`	`1`
`makes me think plurality. Maybe I can love you with many selves. Or. I love all the Porgys. Even as a colloquialism: a queering of love as singular. English is a strange language because I loves and He loves are not both grammarly. I loves you, Porgy. Better to ask what man is not, Porgy. The beauty of Nina’s`	`0`

Loss: BatchAllTripletLoss with these parameters:

{
    "distance_metric": "euclidean_distance",
    "margin": 5
}

Training Hyperparameters

Non-Default Hyperparameters

per_device_train_batch_size: 32
num_train_epochs: 5
learning_rate: 2e-05
warmup_steps: 0.1

All Hyperparameters

Click to expand

per_device_train_batch_size: 32
num_train_epochs: 5
max_steps: -1
learning_rate: 2e-05
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_steps: 0.1
optim: adamw_torch_fused
optim_args: None
weight_decay: 0.0
adam_beta1: 0.9
adam_beta2: 0.999
adam_epsilon: 1e-08
optim_target_modules: None
gradient_accumulation_steps: 1
average_tokens_across_devices: True
max_grad_norm: 1.0
label_smoothing_factor: 0.0
bf16: False
fp16: False
bf16_full_eval: False
fp16_full_eval: False
tf32: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
use_liger_kernel: False
liger_kernel_config: None
use_cache: False
neftune_noise_alpha: None
torch_empty_cache_steps: None
auto_find_batch_size: False
log_on_each_node: True
logging_nan_inf_filter: True
include_num_input_tokens_seen: no
log_level: passive
log_level_replica: warning
disable_tqdm: False
project: huggingface
trackio_space_id: None
trackio_bucket_id: None
trackio_static_space_id: None
per_device_eval_batch_size: 8
prediction_loss_only: True
eval_on_start: False
eval_do_concat_batches: True
eval_use_gather_object: False
eval_accumulation_steps: None
include_for_metrics: []
batch_eval_metrics: False
save_only_model: False
save_on_each_node: False
enable_jit_checkpoint: False
push_to_hub: False
hub_private_repo: None
hub_model_id: None
hub_strategy: every_save
hub_always_push: False
hub_revision: None
load_best_model_at_end: False
ignore_data_skip: False
restore_callback_states_from_checkpoint: False
full_determinism: False
seed: 42
data_seed: None
use_cpu: False
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
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_pin_memory: True
dataloader_persistent_workers: False
dataloader_prefetch_factor: None
remove_unused_columns: True
label_names: None
train_sampling_strategy: random
length_column_name: length
ddp_find_unused_parameters: None
ddp_bucket_cap_mb: None
ddp_broadcast_buffers: False
ddp_static_graph: None
ddp_backend: None
ddp_timeout: 1800
fsdp: None
fsdp_config: None
deepspeed: None
debug: []
skip_memory_metrics: True
do_predict: False
resume_from_checkpoint: None
warmup_ratio: None
local_rank: -1
prompts: None
batch_sampler: batch_sampler
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss
0.3226	10	4.9797
0.6452	20	4.9687
0.9677	30	4.9419
1.2903	40	4.8351
1.6129	50	4.8372
1.9355	60	4.7127
2.2581	70	4.6246
2.5806	80	4.7205
2.9032	90	4.6340
3.2258	100	4.4178
3.5484	110	4.5711
3.8710	120	4.5545
4.1935	130	4.4915
4.5161	140	4.4780
4.8387	150	4.4197

Training Time

Training: 2.4 days

Framework Versions

Python: 3.12.13
Sentence Transformers: 5.6.0
Transformers: 5.12.1
PyTorch: 2.11.0+cpu
Accelerate: 1.14.0
Datasets: 5.0.0
Tokenizers: 0.22.2

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",
}

BatchAllTripletLoss

@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}
}