Classification_Address / README.md

Kao1412

Add new SentenceTransformer model

bafb1de verified 4 months ago

preview code

raw

history blame contribute delete

13.9 kB

metadata

tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - generated_from_trainer
  - dataset_size:6500
  - loss:CosineSimilarityLoss
base_model: keepitreal/vietnamese-sbert
widget:
  - source_sentence: 64 đường tố hữu đông anh hải phòng
    sentences:
      - 64 đường tố hữu đông anh hải phòng
      - 80 mễ trì phú nhuận tp cà mau
      - 81 phùng  khồang  phường  quận  6 đà  nẵng
  - source_sentence: cầu. giấy. chương. mỹ. tphường. hà. tĩnh
    sentences:
      - lê. đức. thọ. hóc. môngõ sóc. trăng
      - phạm hùng phường thanh trì thành phố quy nhơn
      - 74 đường, nguyễn, văn, cừ, phường, thường, tín, đồng, tháp
  - source_sentence: phạm. văngõ bạchuyện đông. anhuyện sóc. trăng
    sentences:
      - số. 95 mễ. trì. phường. hai. bà. trưng. hồ. chí. minh
      - 148 đường trần thái tông bình chánh bình dương
      - số 119 tố hữu tân phú nam định
  - source_sentence: trần thai tong thu đuc soc trầng
    sentences:
      - đức thọ đông anh hải phòng
      - phạm hung quận đan phuong ninh binh
      - trầnthái tông thủ đức sóc trăng
  - source_sentence: xuan, thuy, thanh, tri, đong, thap
    sentences:
      - xuân, thủy, thanh, trì, đồng, tháp
      - so  143 me  tri  quan  10 lam  đong
      - trần  thai  tong  thach  that  thanh  phồ  kon  tum
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy
  - cosine_accuracy_threshold
  - cosine_f1
  - cosine_f1_threshold
  - cosine_precision
  - cosine_recall
  - cosine_ap
  - cosine_mcc
model-index:
  - name: SentenceTransformer based on keepitreal/vietnamese-sbert
    results:
      - task:
          type: binary-classification
          name: Binary Classification
        dataset:
          name: address eval
          type: address-eval
        metrics:
          - type: cosine_accuracy
            value: 0.91
            name: Cosine Accuracy
          - type: cosine_accuracy_threshold
            value: 0.6586315035820007
            name: Cosine Accuracy Threshold
          - type: cosine_f1
            value: 0.9014925373134328
            name: Cosine F1
          - type: cosine_f1_threshold
            value: 0.6586315035820007
            name: Cosine F1 Threshold
          - type: cosine_precision
            value: 0.897029702970297
            name: Cosine Precision
          - type: cosine_recall
            value: 0.906
            name: Cosine Recall
          - type: cosine_ap
            value: 0.9161149688703704
            name: Cosine Ap
          - type: cosine_mcc
            value: 0.8186854636882175
            name: Cosine Mcc

SentenceTransformer based on keepitreal/vietnamese-sbert

This is a sentence-transformers model finetuned from keepitreal/vietnamese-sbert. 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: keepitreal/vietnamese-sbert
Maximum Sequence Length: 256 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': 256, 'do_lower_case': False}) with Transformer model: RobertaModel 
  (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("Kao1412/Classification_Address")
# Run inference
sentences = [
    'xuan, thuy, thanh, tri, đong, thap',
    'xuân, thủy, thanh, trì, đồng, tháp',
    'trần  thai  tong  thach  that  thanh  phồ  kon  tum',
]
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

Binary Classification

Dataset: address-eval
Evaluated with BinaryClassificationEvaluator

Metric	Value
cosine_accuracy	0.91
cosine_accuracy_threshold	0.6586
cosine_f1	0.9015
cosine_f1_threshold	0.6586
cosine_precision	0.897
cosine_recall	0.906
cosine_ap	0.9161
cosine_mcc	0.8187

Training Details

Training Dataset

Unnamed Dataset

Size: 6,500 training samples
Columns: sentence_0, sentence_1, and label
Approximate statistics based on the first 1000 samples:
sentence_0 sentence_1 label
type string string float
details
min: 4 tokens
mean: 14.83 tokens
max: 34 tokens

min: 4 tokens
mean: 14.54 tokens
max: 34 tokens

min: 0.0
mean: 0.47
max: 1.0

	sentence_0	sentence_1	label
type	string	string	float
details	min: 4 tokens mean: 14.83 tokens max: 34 tokens	min: 4 tokens mean: 14.54 tokens max: 34 tokens	min: 0.0 mean: 0.47 max: 1.0

Samples:

sentence_0	sentence_1	label
`42 lê van lương ba đình an giang`	`42 lê văn lương ba đình an giang`	`1.0`
`so 51 đuong nguyễn chi thanh đong anh đa nang`	`phạm van bach phu xuyen soc trầng`	`0.0`
`phồ, le, van, luong, phu, nhuan, long, an`	`phồ, le, văn, lương, phu, nhuan, long, an`	`1.0`

Loss: CosineSimilarityLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

eval_strategy: steps
per_device_train_batch_size: 32
per_device_eval_batch_size: 32
num_train_epochs: 5
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: 32
per_device_eval_batch_size: 32
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: 5
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: 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
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	Training Loss	address-eval_cosine_ap
1.0	204	-	0.8984
2.0	408	-	0.9073
2.4510	500	0.0884	0.9108
3.0	612	-	0.9118
4.0	816	-	0.9147
4.9020	1000	0.0627	0.9161

Framework Versions

Python: 3.11.12
Sentence Transformers: 4.1.0
Transformers: 4.52.3
PyTorch: 2.6.0+cu124
Accelerate: 1.7.0
Datasets: 2.14.4
Tokenizers: 0.21.1

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",
}