Tay_Embedding / README.md

Tay–Vietnamese embedding trained with contrastive learning

180af21 verified 12 days ago

16.7 kB

	---
	tags:
	- sentence-transformers
	- sentence-similarity
	- feature-extraction
	- dense
	- generated_from_trainer
	- dataset_size:20554
	- loss:MultipleNegativesRankingLoss
	base_model: AITeamVN/Vietnamese_Embedding_v2
	widget:
	- source_sentence: bon
	sentences:
	- cây mon
	- đổ chậu nước
	- yên phận làm ăn
	- source_sentence: Tua cáy chọt oóc khói doòng
	sentences:
	- chăn
	- hen thở khò khè
	- con gà xổng ra khỏi lồng
	- source_sentence: Khảm
	sentences:
	- kiểm tra
	- treo
	- rạo rực
	- source_sentence: khẩu hảo Bẩu
	sentences:
	- mẹ mắng không bằng bố sa sầm mặt
	- cạo trọc đầu
	- thóc chưa khô hẳn
	- source_sentence: Các
	sentences:
	- mập mạp
	- chân tay mập
	- bắc
	datasets:
	- HeyDunaX/tay-vietnamese-nmt
	pipeline_tag: sentence-similarity
	library_name: sentence-transformers
	---

	# SentenceTransformer based on AITeamVN/Vietnamese_Embedding_v2

	This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [AITeamVN/Vietnamese_Embedding_v2](https://huggingface.co/AITeamVN/Vietnamese_Embedding_v2) on the [tay-vietnamese-nmt](https://huggingface.co/datasets/HeyDunaX/tay-vietnamese-nmt) dataset. It maps sentences & paragraphs to a 1024-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: [AITeamVN/Vietnamese_Embedding_v2](https://huggingface.co/AITeamVN/Vietnamese_Embedding_v2) <!-- at revision 18b44161e041bf1d3a333ab5144b5b7b93f914d2 -->
	- Maximum Sequence Length: 8192 tokens
	- Output Dimensionality: 1024 dimensions
	- Similarity Function: Cosine Similarity
	- Training Dataset:
	- [tay-vietnamese-nmt](https://huggingface.co/datasets/HeyDunaX/tay-vietnamese-nmt)
	<!-- - Language: Unknown -->
	<!-- - License: Unknown -->

	### Model Sources

	- Documentation: [Sentence Transformers Documentation](https://sbert.net)
	- Repository: [Sentence Transformers on GitHub](https://github.com/huggingface/sentence-transformers)
	- Hugging Face: [Sentence Transformers on Hugging Face](https://huggingface.co/models?library=sentence-transformers)

	### Full Model Architecture

	```
	SentenceTransformer(
	(0): Transformer({'max_seq_length': 8192, 'do_lower_case': False, 'architecture': 'XLMRobertaModel'})
	(1): Pooling({'word_embedding_dimension': 1024, '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:

	```bash
	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("HeyDunaX/Tay_Embedding")
	# Run inference
	sentences = [
	'Các',
	'bắc',
	'chân tay mập',
	]
	embeddings = model.encode(sentences)
	print(embeddings.shape)
	# [3, 1024]

	# Get the similarity scores for the embeddings
	similarities = model.similarity(embeddings, embeddings)
	print(similarities)
	# tensor([[ 1.0000, 0.3147, -0.0254],
	# [ 0.3147, 1.0000, -0.1489],
	# [-0.0254, -0.1489, 1.0000]])
	```

	<!--
	### Direct Usage (Transformers)

	<details><summary>Click to see the direct usage in Transformers</summary>

	</details>
	-->

	<!--
	### Downstream Usage (Sentence Transformers)

	You can finetune this model on your own dataset.

	<details><summary>Click to expand</summary>

	</details>
	-->

	<!--
	### Out-of-Scope Use

	List how the model may foreseeably be misused and address what users ought not to do with the model.
	-->

	<!--
	## Bias, Risks and Limitations

	What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model.
	-->

	<!--
	### Recommendations

	What are recommendations with respect to the foreseeable issues? For example, filtering explicit content.
	-->

	## Training Details

	### Training Dataset

	#### tay-vietnamese-nmt

	* Dataset: [tay-vietnamese-nmt](https://huggingface.co/datasets/HeyDunaX/tay-vietnamese-nmt) at [2b04e13](https://huggingface.co/datasets/HeyDunaX/tay-vietnamese-nmt/tree/2b04e139b670d8ad62693d1fbdb943940e4acc05)
	* Size: 20,554 training samples
	* Columns: <code>sentence1</code> and <code>sentence2</code>
	* Approximate statistics based on the first 1000 samples:
	\| \| sentence1 \| sentence2 \|
	\|:--------\|:---------------------------------------------------------------------------------\|:---------------------------------------------------------------------------------\|
	\| type \| string \| string \|
	\| details \| <ul><li>min: 3 tokens</li><li>mean: 6.77 tokens</li><li>max: 21 tokens</li></ul> \| <ul><li>min: 3 tokens</li><li>mean: 5.85 tokens</li><li>max: 17 tokens</li></ul> \|
	* Samples:
	\| sentence1 \| sentence2 \|
	\|:--------------------------\|:------------------------\|
	\| <code>me</code> \| <code>bà cô</code> \|
	\| <code>noọng ấc cải</code> \| <code>em ngực bự</code> \|
	\| <code>noọng</code> \| <code>em gái</code> \|
	* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
	```json
	{
	"scale": 20.0,
	"similarity_fct": "cos_sim",
	"gather_across_devices": false
	}
	```

	### Evaluation Dataset

	#### tay-vietnamese-nmt

	* Dataset: [tay-vietnamese-nmt](https://huggingface.co/datasets/HeyDunaX/tay-vietnamese-nmt) at [2b04e13](https://huggingface.co/datasets/HeyDunaX/tay-vietnamese-nmt/tree/2b04e139b670d8ad62693d1fbdb943940e4acc05)
	* Size: 2,295 evaluation samples
	* Columns: <code>sentence1</code> and <code>sentence2</code>
	* Approximate statistics based on the first 1000 samples:
	\| \| sentence1 \| sentence2 \|
	\|:--------\|:---------------------------------------------------------------------------------\|:---------------------------------------------------------------------------------\|
	\| type \| string \| string \|
	\| details \| <ul><li>min: 3 tokens</li><li>mean: 7.24 tokens</li><li>max: 26 tokens</li></ul> \| <ul><li>min: 3 tokens</li><li>mean: 6.02 tokens</li><li>max: 22 tokens</li></ul> \|
	* Samples:
	\| sentence1 \| sentence2 \|
	\|:-------------------------------------\|:--------------------------------------------\|
	\| <code>Hết fiệc ác</code> \| <code>làm việc khoẻ</code> \|
	\| <code>slấc ác</code> \| <code>giặc độc ác</code> \|
	\| <code>ái chin mác rèo năm mạy</code> \| <code>Muốn ăn quả thì phải trồng cây</code> \|
	* Loss: [<code>MultipleNegativesRankingLoss</code>](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#multiplenegativesrankingloss) with these parameters:
	```json
	{
	"scale": 20.0,
	"similarity_fct": "cos_sim",
	"gather_across_devices": false
	}
	```

	### Training Hyperparameters
	#### Non-Default Hyperparameters

	- `eval_strategy`: epoch
	- `gradient_accumulation_steps`: 4
	- `learning_rate`: 1e-05
	- `num_train_epochs`: 10
	- `warmup_ratio`: 0.1
	- `warmup_steps`: 0.1
	- `fp16`: True
	- `load_best_model_at_end`: True

	#### All Hyperparameters
	<details><summary>Click to expand</summary>

	- `do_predict`: False
	- `eval_strategy`: epoch
	- `prediction_loss_only`: True
	- `per_device_train_batch_size`: 8
	- `per_device_eval_batch_size`: 8
	- `gradient_accumulation_steps`: 4
	- `eval_accumulation_steps`: None
	- `torch_empty_cache_steps`: None
	- `learning_rate`: 1e-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`: 10
	- `max_steps`: -1
	- `lr_scheduler_type`: linear
	- `lr_scheduler_kwargs`: None
	- `warmup_ratio`: 0.1
	- `warmup_steps`: 0.1
	- `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`: True
	- `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`: batch_sampler
	- `multi_dataset_batch_sampler`: proportional
	- `router_mapping`: {}
	- `learning_rate_mapping`: {}

	</details>

	### Training Logs
	\| Epoch \| Step \| Training Loss \| Validation Loss \|
	\|:------:\|:----:\|:-------------:\|:---------------:\|
	\| 0.1556 \| 100 \| 1.7414 \| - \|
	\| 0.3113 \| 200 \| 1.3566 \| - \|
	\| 0.4669 \| 300 \| 1.1332 \| - \|
	\| 0.6226 \| 400 \| 1.0198 \| - \|
	\| 0.7782 \| 500 \| 0.8943 \| - \|
	\| 0.9339 \| 600 \| 0.7909 \| - \|
	\| 1.0 \| 643 \| - \| 0.7135 \|
	\| 1.0887 \| 700 \| 0.7070 \| - \|
	\| 1.2444 \| 800 \| 0.6029 \| - \|
	\| 1.4 \| 900 \| 0.6095 \| - \|
	\| 1.5556 \| 1000 \| 0.5436 \| - \|
	\| 1.7113 \| 1100 \| 0.5534 \| - \|
	\| 1.8669 \| 1200 \| 0.5363 \| - \|
	\| 2.0 \| 1286 \| - \| 0.5121 \|
	\| 2.0218 \| 1300 \| 0.4886 \| - \|
	\| 2.1774 \| 1400 \| 0.3853 \| - \|
	\| 2.3331 \| 1500 \| 0.3940 \| - \|
	\| 2.4887 \| 1600 \| 0.3859 \| - \|
	\| 2.6444 \| 1700 \| 0.4035 \| - \|
	\| 2.8 \| 1800 \| 0.3686 \| - \|
	\| 2.9556 \| 1900 \| 0.3662 \| - \|
	\| 3.0 \| 1929 \| - \| 0.4505 \|
	\| 3.1105 \| 2000 \| 0.3276 \| - \|
	\| 3.2661 \| 2100 \| 0.2877 \| - \|
	\| 3.4218 \| 2200 \| 0.2991 \| - \|
	\| 3.5774 \| 2300 \| 0.2898 \| - \|
	\| 3.7331 \| 2400 \| 0.2704 \| - \|
	\| 3.8887 \| 2500 \| 0.2807 \| - \|
	\| 4.0 \| 2572 \| - \| 0.4247 \|
	\| 4.0436 \| 2600 \| 0.2879 \| - \|
	\| 4.1992 \| 2700 \| 0.2300 \| - \|
	\| 4.3549 \| 2800 \| 0.2233 \| - \|
	\| 4.5105 \| 2900 \| 0.2169 \| - \|
	\| 4.6661 \| 3000 \| 0.2273 \| - \|
	\| 4.8218 \| 3100 \| 0.2149 \| - \|
	\| 4.9774 \| 3200 \| 0.2277 \| - \|
	\| 5.0 \| 3215 \| - \| 0.4163 \|
	\| 5.1323 \| 3300 \| 0.1973 \| - \|
	\| 5.2879 \| 3400 \| 0.1856 \| - \|
	\| 5.4436 \| 3500 \| 0.1686 \| - \|
	\| 5.5992 \| 3600 \| 0.1797 \| - \|
	\| 5.7549 \| 3700 \| 0.1830 \| - \|
	\| 5.9105 \| 3800 \| 0.1701 \| - \|
	\| 6.0 \| 3858 \| - \| 0.4066 \|
	\| 6.0654 \| 3900 \| 0.1620 \| - \|
	\| 6.2210 \| 4000 \| 0.1453 \| - \|
	\| 6.3767 \| 4100 \| 0.1593 \| - \|
	\| 6.5323 \| 4200 \| 0.1481 \| - \|
	\| 6.6879 \| 4300 \| 0.1506 \| - \|
	\| 6.8436 \| 4400 \| 0.1534 \| - \|
	\| 6.9992 \| 4500 \| 0.1554 \| - \|
	\| 7.0 \| 4501 \| - \| 0.3907 \|
	\| 7.1541 \| 4600 \| 0.1284 \| - \|
	\| 7.3097 \| 4700 \| 0.1266 \| - \|
	\| 7.4654 \| 4800 \| 0.1392 \| - \|
	\| 7.6210 \| 4900 \| 0.1292 \| - \|
	\| 7.7767 \| 5000 \| 0.1309 \| - \|
	\| 7.9323 \| 5100 \| 0.1318 \| - \|
	\| 8.0 \| 5144 \| - \| 0.3922 \|
	\| 8.0872 \| 5200 \| 0.1263 \| - \|
	\| 8.2428 \| 5300 \| 0.1136 \| - \|
	\| 8.3984 \| 5400 \| 0.1161 \| - \|
	\| 8.5541 \| 5500 \| 0.1137 \| - \|
	\| 8.7097 \| 5600 \| 0.1231 \| - \|
	\| 8.8654 \| 5700 \| 0.1187 \| - \|
	\| 9.0 \| 5787 \| - \| 0.3875 \|
	\| 9.0202 \| 5800 \| 0.1182 \| - \|
	\| 9.1759 \| 5900 \| 0.1059 \| - \|
	\| 9.3315 \| 6000 \| 0.1062 \| - \|
	\| 9.4872 \| 6100 \| 0.1044 \| - \|
	\| 9.6428 \| 6200 \| 0.0992 \| - \|
	\| 9.7984 \| 6300 \| 0.1057 \| - \|
	\| 9.9541 \| 6400 \| 0.1048 \| - \|
	\| 10.0 \| 6430 \| - \| 0.3878 \|


	### Framework Versions
	- Python: 3.12.12
	- Sentence Transformers: 5.2.2
	- Transformers: 5.0.0
	- PyTorch: 2.9.0+cu126
	- Accelerate: 1.12.0
	- Datasets: 4.0.0
	- Tokenizers: 0.22.2

	## 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",
	}
	```

	#### MultipleNegativesRankingLoss
	```bibtex
	@misc{henderson2017efficient,
	title={Efficient Natural Language Response Suggestion for Smart Reply},
	author={Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-hsuan Sung and Laszlo Lukacs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
	year={2017},
	eprint={1705.00652},
	archivePrefix={arXiv},
	primaryClass={cs.CL}
	}
	```

	<!--
	## Glossary

	Clearly define terms in order to be accessible across audiences.
	-->

	<!--
	## Model Card Authors

	Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction.
	-->

	<!--
	## Model Card Contact

	Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors.
	-->