Add new SentenceTransformer model

605ad24 verified 6 days ago

17 kB

language:
  - en
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - dense
  - generated_from_trainer
  - dataset_size:3000
  - loss:BatchAllTripletLoss
base_model: microsoft/mpnet-base
widget:
  - source_sentence: what am i supposed to do if i lost my luggage
    sentences:
      - do i need a visa if i go there
      - why did you freeze my bank account
      - tell my bank that i'm travelling to france in 2 days
  - source_sentence: can you suggest some of the most popular travel destination
    sentences:
      - what is the total of my repair bill
      - could you tell me my bill's minimum payment
      - >-
        can you get me a car rental for march 1st to 3rd in seattle, and i'd
        like a sedan if possible
  - source_sentence: is there a minimum amount accepted
    sentences:
      - am i going to need a visa for traveling to canada
      - submit payment to duke energy for my electric bill
      - let me know chase's routing number
  - source_sentence: my account appears to be blocked and i don't know why
    sentences:
      - how do you say hello in japanese
      - how much is due on the gas bill
      - how much was my last transaction for
  - source_sentence: are there any travel alerts for juarez
    sentences:
      - i am now out of checks, how do i order new ones
      - lowest amount for cable bill
      - how much interest do i get on my citizen's savings account
datasets:
  - contemmcm/clinc150
pipeline_tag: sentence-similarity
library_name: sentence-transformers

SentenceTransformer based on microsoft/mpnet-base

This is a sentence-transformers model finetuned from microsoft/mpnet-base on the clinc150 dataset. 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: microsoft/mpnet-base
Maximum Sequence Length: 512 tokens
Output Dimensionality: 768 dimensions
Similarity Function: Cosine Similarity
Training Dataset:
- clinc150
Language: en

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, 'architecture': '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})
)

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("bnoland/mpnet-base-clinc-subset")
# Run inference
sentences = [
    'are there any travel alerts for juarez',
    "how much interest do i get on my citizen's savings account",
    'lowest amount for cable bill',
]
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, 0.7056, 0.6717],
#         [0.7056, 1.0000, 0.7377],
#         [0.6717, 0.7377, 1.0000]])

Training Details

Training Dataset

clinc150

Dataset: clinc150 at 2bbb9af
Size: 3,000 training samples
Columns: text and label

Approximate statistics based on the first 1000 samples:

	text	label
type	string	int
details	min: 6 tokens mean: 12.61 tokens max: 28 tokens	1: ~3.60% 2: ~3.80% 3: ~3.50% 4: ~4.30% 5: ~3.90% 6: ~3.50% 7: ~2.20% 8: ~3.00% 9: ~2.80% 10: ~2.90% 11: ~3.70% 12: ~2.80% 13: ~3.70% 14: ~2.80% 15: ~3.90% 76: ~3.60% 77: ~3.40% 78: ~3.60% 79: ~3.40% 80: ~3.20% 81: ~3.70% 82: ~3.00% 83: ~2.90% 84: ~3.30% 85: ~3.50% 86: ~3.70% 87: ~2.40% 88: ~3.70% 89: ~2.70% 90: ~3.50%

Samples:

text label

is there enough money in my bank of hawaii for vacation 12

i need to let my bank know i am visiting asia soon 77

what's bank of america's routing number 2
Loss: BatchAllTripletLoss

text	label
`is there enough money in my bank of hawaii for vacation`	`12`
`i need to let my bank know i am visiting asia soon`	`77`
`what's bank of america's routing number`	`2`

Evaluation Dataset

clinc150

Dataset: clinc150 at 2bbb9af
Size: 600 evaluation samples
Columns: text and label

Approximate statistics based on the first 600 samples:

	text	label
type	string	int
details	min: 6 tokens mean: 12.83 tokens max: 26 tokens	1: ~3.33% 2: ~3.33% 3: ~3.33% 4: ~3.33% 5: ~3.33% 6: ~3.33% 7: ~3.33% 8: ~3.33% 9: ~3.33% 10: ~3.33% 11: ~3.33% 12: ~3.33% 13: ~3.33% 14: ~3.33% 15: ~3.33% 76: ~3.33% 77: ~3.33% 78: ~3.33% 79: ~3.33% 80: ~3.33% 81: ~3.33% 82: ~3.33% 83: ~3.33% 84: ~3.33% 85: ~3.33% 86: ~3.33% 87: ~3.33% 88: ~3.33% 89: ~3.33% 90: ~3.33%

Samples:

text label

was my last transaction at walmart 14

what interest rate is us bank giving me on my acount 7

look up carry-on rules for american airlines 89
Loss: BatchAllTripletLoss

text	label
`was my last transaction at walmart`	`14`
`what interest rate is us bank giving me on my acount`	`7`
`look up carry-on rules for american airlines`	`89`

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
num_train_epochs: 1
warmup_steps: 10
fp16: True
batch_sampler: group_by_label

All Hyperparameters

Click to expand

do_predict: False
eval_strategy: steps
prediction_loss_only: True
per_device_train_batch_size: 16
per_device_eval_batch_size: 16
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: 1
max_steps: -1
lr_scheduler_type: linear
lr_scheduler_kwargs: None
warmup_ratio: None
warmup_steps: 10
log_level: passive
log_level_replica: warning
log_on_each_node: True
logging_nan_inf_filter: True
enable_jit_checkpoint: False
save_on_each_node: False
save_only_model: False
restore_callback_states_from_checkpoint: False
use_cpu: False
seed: 42
data_seed: None
bf16: False
fp16: True
bf16_full_eval: False
fp16_full_eval: False
tf32: None
local_rank: -1
ddp_backend: None
debug: []
dataloader_drop_last: False
dataloader_num_workers: 0
dataloader_prefetch_factor: None
disable_tqdm: False
remove_unused_columns: True
label_names: None
load_best_model_at_end: False
ignore_data_skip: False
fsdp: []
fsdp_config: {'min_num_params': 0, 'xla': False, 'xla_fsdp_v2': False, 'xla_fsdp_grad_ckpt': 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
deepspeed: None
label_smoothing_factor: 0.0
optim: adamw_torch_fused
optim_args: None
group_by_length: False
length_column_name: length
project: huggingface
trackio_space_id: trackio
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
push_to_hub: False
resume_from_checkpoint: None
hub_model_id: None
hub_strategy: every_save
hub_private_repo: None
hub_always_push: False
hub_revision: None
gradient_checkpointing: False
gradient_checkpointing_kwargs: None
include_for_metrics: []
eval_do_concat_batches: True
auto_find_batch_size: False
full_determinism: False
ddp_timeout: 1800
torch_compile: False
torch_compile_backend: None
torch_compile_mode: None
include_num_input_tokens_seen: no
neftune_noise_alpha: None
optim_target_modules: None
batch_eval_metrics: False
eval_on_start: False
use_liger_kernel: False
liger_kernel_config: None
eval_use_gather_object: False
average_tokens_across_devices: True
use_cache: False
prompts: None
batch_sampler: group_by_label
multi_dataset_batch_sampler: proportional
router_mapping: {}
learning_rate_mapping: {}

Training Logs

Epoch	Step	Training Loss	Validation Loss
0.5319	100	0.5093	1.7369

Framework Versions

Python: 3.12.12
Sentence Transformers: 5.2.3
Transformers: 5.0.0
PyTorch: 2.10.0+cu128
Accelerate: 1.12.0
Datasets: 4.0.0
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}
}