intent/m3e

---

datasets: [] language: [] library_name: sentence-transformers pipeline_tag: sentence-similarity tags:

• sentence-transformers
• sentence-similarity
• feature-extraction
• generated_from_trainer
• dataset_size:588644
• loss:CosineSimilarityLoss widget:
• source_sentence: 我这水表冻裂了，寻思换一个多钱啊 sentences:
  • 需本人到供水公司，携带“用户申请报停审批表”进行办理。
  • 供水营业厅缴费地址：一、供水客户服务中心：会展中心道东光复西路790号。 二、佳东供水收费厅:长胜路与光复路交叉口长胜街85号。三、九小供水收费厅：第九小区大门西侧路北兴城胡同128号。四、行政服务中心一楼：长安路西段820号
  • 线上报装登录佳木斯供水公众号， 点击“报装申请”，提交后，工作人员会在24小时内联系您。线下报装需本人到供水公司申请，居民用户携带本人身份证复印件和房照复印件，非居民用户携带营业执照复印件、法人复印件及公章。
• source_sentence: 申请安装自来水需要哪些具体条件？ sentences:
  • 怎么停水了？
  • 您可以登录微信公众号佳木斯供水，进入主页面绑定用户信息后缴费，或者微信小程序中缴费页面输入用户编号进行缴费。或者也可以在支付宝生活缴费页面输入用户编号进行缴费。
  • 不是
• source_sentence: 咱家水表不好使，屋里头水都关了，不用水，它自己走 sentences:
  • 缴水费后需要联系抄表员后，24小时内给水。
  • 您可以登录微信公众号佳木斯供水，进入主页面绑定用户信息后缴费，或者微信小程序中缴费页面输入用户编号进行缴费。或者也可以在支付宝生活缴费页面输入用户编号进行缴费。
  • 怎么停水了？
• source_sentence: 我想咨询个事儿，我家交水费这个名，后面有一个字错了，怎么改一下子 sentences:
  • 怎么停水了？
  • 我家漏水了
  • 缴水费后需要联系抄表员后，24小时内给水。
• source_sentence: 我家厨房的管可能冻了，没有水 sentences:
  • 怎么停水了？
  • 您可以登录微信公众号佳木斯供水，进入主页面绑定用户信息后缴费，或者微信小程序中缴费页面输入用户编号进行缴费。或者也可以在支付宝生活缴费页面输入用户编号进行缴费。
  • 水表坏了 ---

SentenceTransformer

This is a sentence-transformers model trained. 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
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• 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': False}) with Transformer model: BertModel 
  (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("sentence_transformers_model_id")
# Run inference
sentences = [
    '我家厨房的管可能冻了，没有水',
    '怎么停水了？',
    '水表坏了',
]
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]

Training Details

Training Dataset

Unnamed Dataset

• Size: 588,644 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: 3 tokens
  • mean: 14.9 tokens
  • max: 57 tokens
  |
  • min: 4 tokens
  • mean: 23.63 tokens
  • max: 107 tokens
  |
  • min: 0.3
  • mean: 0.3
  • max: 0.7
  |
• Samples: | sentence_0 | sentence_1 | label | |:---------------------------------------|:--------------------|:-----------------| | 我想问一下我家，三号楼五单元，为什么那个停水啊？ | 阀门问题 | 0.3 | | 你好，刚才我打电话了，那个用户更名都用什么证件呢？ | 怎么停水了？ | 0.3 | | 再说一下 | 怎么停水了？ | 0.3 |

• Loss: CosineSimilarityLoss with these parameters:

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

Training Hyperparameters

Non-Default Hyperparameters

• per_device_train_batch_size: 16
• per_device_eval_batch_size: 16
• multi_dataset_batch_sampler: round_robin

All Hyperparameters

Click to expand

- `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: no - `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`: 3 - `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`: False - `hub_always_push`: False - `gradient_checkpointing`: False - `gradient_checkpointing_kwargs`: None - `include_inputs_for_metrics`: False - `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 - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: round_robin

Training Logs

Click to expand

| Epoch | Step | Training Loss | |:------:|:------:|:-------------:| | 0.0136 | 500 | 0.21 | | 0.0272 | 1000 | 0.1795 | | 0.0408 | 1500 | 0.1296 | | 0.0544 | 2000 | 0.0749 | | 0.0680 | 2500 | 0.0312 | | 0.0815 | 3000 | 0.0137 | | 0.0951 | 3500 | 0.0079 | | 0.1087 | 4000 | 0.0057 | | 0.1223 | 4500 | 0.0042 | | 0.1359 | 5000 | 0.0038 | | 0.1495 | 5500 | 0.0033 | | 0.1631 | 6000 | 0.0029 | | 0.1767 | 6500 | 0.0028 | | 0.1903 | 7000 | 0.0026 | | 0.2039 | 7500 | 0.0024 | | 0.2174 | 8000 | 0.0022 | | 0.2310 | 8500 | 0.0019 | | 0.2446 | 9000 | 0.0019 | | 0.2582 | 9500 | 0.002 | | 0.2718 | 10000 | 0.0019 | | 0.2854 | 10500 | 0.0019 | | 0.2990 | 11000 | 0.0017 | | 0.3126 | 11500 | 0.0016 | | 0.3262 | 12000 | 0.0016 | | 0.3398 | 12500 | 0.0016 | | 0.3533 | 13000 | 0.0015 | | 0.3669 | 13500 | 0.0014 | | 0.3805 | 14000 | 0.0013 | | 0.3941 | 14500 | 0.0015 | | 0.4077 | 15000 | 0.0014 | | 0.4213 | 15500 | 0.0012 | | 0.4349 | 16000 | 0.0013 | | 0.4485 | 16500 | 0.0012 | | 0.4621 | 17000 | 0.0013 | | 0.4757 | 17500 | 0.0012 | | 0.4893 | 18000 | 0.0011 | | 0.5028 | 18500 | 0.0011 | | 0.5164 | 19000 | 0.0011 | | 0.5300 | 19500 | 0.0012 | | 0.5436 | 20000 | 0.001 | | 0.5572 | 20500 | 0.0012 | | 0.5708 | 21000 | 0.0009 | | 0.5844 | 21500 | 0.0009 | | 0.5980 | 22000 | 0.0008 | | 0.6116 | 22500 | 0.0009 | | 0.6252 | 23000 | 0.0009 | | 0.6387 | 23500 | 0.0008 | | 0.6523 | 24000 | 0.0007 | | 0.6659 | 24500 | 0.0008 | | 0.6795 | 25000 | 0.0009 | | 0.6931 | 25500 | 0.0007 | | 0.7067 | 26000 | 0.0008 | | 0.7203 | 26500 | 0.0008 | | 0.7339 | 27000 | 0.0007 | | 0.7475 | 27500 | 0.0006 | | 0.7611 | 28000 | 0.0006 | | 0.7746 | 28500 | 0.0006 | | 0.7882 | 29000 | 0.0006 | | 0.8018 | 29500 | 0.0007 | | 0.8154 | 30000 | 0.0006 | | 0.8290 | 30500 | 0.0007 | | 0.8426 | 31000 | 0.0007 | | 0.8562 | 31500 | 0.0006 | | 0.8698 | 32000 | 0.0006 | | 0.8834 | 32500 | 0.0006 | | 0.8970 | 33000 | 0.0006 | | 0.9105 | 33500 | 0.0007 | | 0.9241 | 34000 | 0.0005 | | 0.9377 | 34500 | 0.0007 | | 0.9513 | 35000 | 0.0006 | | 0.9649 | 35500 | 0.0006 | | 0.9785 | 36000 | 0.0006 | | 0.9921 | 36500 | 0.0005 | | 1.0057 | 37000 | 0.0004 | | 1.0193 | 37500 | 0.0005 | | 1.0329 | 38000 | 0.0005 | | 1.0465 | 38500 | 0.0006 | | 1.0600 | 39000 | 0.0005 | | 1.0736 | 39500 | 0.0005 | | 1.0872 | 40000 | 0.0005 | | 1.1008 | 40500 | 0.0005 | | 1.1144 | 41000 | 0.0006 | | 1.1280 | 41500 | 0.0004 | | 1.1416 | 42000 | 0.0005 | | 1.1552 | 42500 | 0.0004 | | 1.1688 | 43000 | 0.0005 | | 1.1824 | 43500 | 0.0004 | | 1.1959 | 44000 | 0.0004 | | 1.2095 | 44500 | 0.0005 | | 1.2231 | 45000 | 0.0004 | | 1.2367 | 45500 | 0.0004 | | 1.2503 | 46000 | 0.0004 | | 1.2639 | 46500 | 0.0004 | | 1.2775 | 47000 | 0.0004 | | 1.2911 | 47500 | 0.0003 | | 1.3047 | 48000 | 0.0004 | | 1.3183 | 48500 | 0.0004 | | 1.3318 | 49000 | 0.0003 | | 1.3454 | 49500 | 0.0004 | | 1.3590 | 50000 | 0.0004 | | 1.3726 | 50500 | 0.0002 | | 1.3862 | 51000 | 0.0003 | | 1.3998 | 51500 | 0.0004 | | 1.4134 | 52000 | 0.0004 | | 1.4270 | 52500 | 0.0003 | | 1.4406 | 53000 | 0.0003 | | 1.4542 | 53500 | 0.0003 | | 1.4678 | 54000 | 0.0005 | | 1.4813 | 54500 | 0.0003 | | 1.4949 | 55000 | 0.0002 | | 1.5085 | 55500 | 0.0003 | | 1.5221 | 56000 | 0.0004 | | 1.5357 | 56500 | 0.0004 | | 1.5493 | 57000 | 0.0004 | | 1.5629 | 57500 | 0.0004 | | 1.5765 | 58000 | 0.0002 | | 1.5901 | 58500 | 0.0003 | | 1.6037 | 59000 | 0.0002 | | 1.6172 | 59500 | 0.0003 | | 1.6308 | 60000 | 0.0003 | | 1.6444 | 60500 | 0.0003 | | 1.6580 | 61000 | 0.0002 | | 1.6716 | 61500 | 0.0004 | | 1.6852 | 62000 | 0.0004 | | 1.6988 | 62500 | 0.0003 | | 1.7124 | 63000 | 0.0003 | | 1.7260 | 63500 | 0.0002 | | 1.7396 | 64000 | 0.0003 | | 1.7531 | 64500 | 0.0002 | | 1.7667 | 65000 | 0.0002 | | 1.7803 | 65500 | 0.0002 | | 1.7939 | 66000 | 0.0003 | | 1.8075 | 66500 | 0.0003 | | 1.8211 | 67000 | 0.0003 | | 1.8347 | 67500 | 0.0003 | | 1.8483 | 68000 | 0.0003 | | 1.8619 | 68500 | 0.0002 | | 1.8755 | 69000 | 0.0003 | | 1.8890 | 69500 | 0.0003 | | 1.9026 | 70000 | 0.0003 | | 1.9162 | 70500 | 0.0002 | | 1.9298 | 71000 | 0.0003 | | 1.9434 | 71500 | 0.0002 | | 1.9570 | 72000 | 0.0003 | | 1.9706 | 72500 | 0.0003 | | 1.9842 | 73000 | 0.0002 | | 1.9978 | 73500 | 0.0002 | | 2.0114 | 74000 | 0.0003 | | 2.0250 | 74500 | 0.0002 | | 2.0385 | 75000 | 0.0002 | | 2.0521 | 75500 | 0.0002 | | 2.0657 | 76000 | 0.0003 | | 2.0793 | 76500 | 0.0002 | | 2.0929 | 77000 | 0.0001 | | 2.1065 | 77500 | 0.0002 | | 2.1201 | 78000 | 0.0002 | | 2.1337 | 78500 | 0.0003 | | 2.1473 | 79000 | 0.0002 | | 2.1609 | 79500 | 0.0003 | | 2.1744 | 80000 | 0.0002 | | 2.1880 | 80500 | 0.0002 | | 2.2016 | 81000 | 0.0002 | | 2.2152 | 81500 | 0.0002 | | 2.2288 | 82000 | 0.0002 | | 2.2424 | 82500 | 0.0002 | | 2.2560 | 83000 | 0.0002 | | 2.2696 | 83500 | 0.0002 | | 2.2832 | 84000 | 0.0002 | | 2.2968 | 84500 | 0.0002 | | 2.3103 | 85000 | 0.0002 | | 2.3239 | 85500 | 0.0004 | | 2.3375 | 86000 | 0.0002 | | 2.3511 | 86500 | 0.0001 | | 2.3647 | 87000 | 0.0003 | | 2.3783 | 87500 | 0.0001 | | 2.3919 | 88000 | 0.0002 | | 2.4055 | 88500 | 0.0002 | | 2.4191 | 89000 | 0.0002 | | 2.4327 | 89500 | 0.0002 | | 2.4463 | 90000 | 0.0001 | | 2.4598 | 90500 | 0.0002 | | 2.4734 | 91000 | 0.0002 | | 2.4870 | 91500 | 0.0002 | | 2.5006 | 92000 | 0.0002 | | 2.5142 | 92500 | 0.0002 | | 2.5278 | 93000 | 0.0002 | | 2.5414 | 93500 | 0.0003 | | 2.5550 | 94000 | 0.0003 | | 2.5686 | 94500 | 0.0002 | | 2.5822 | 95000 | 0.0002 | | 2.5957 | 95500 | 0.0002 | | 2.6093 | 96000 | 0.0002 | | 2.6229 | 96500 | 0.0001 | | 2.6365 | 97000 | 0.0002 | | 2.6501 | 97500 | 0.0001 | | 2.6637 | 98000 | 0.0003 | | 2.6773 | 98500 | 0.0002 | | 2.6909 | 99000 | 0.0002 | | 2.7045 | 99500 | 0.0002 | | 2.7181 | 100000 | 0.0002 | | 2.7316 | 100500 | 0.0002 | | 2.7452 | 101000 | 0.0001 | | 2.7588 | 101500 | 0.0001 | | 2.7724 | 102000 | 0.0002 | | 2.7860 | 102500 | 0.0001 | | 2.7996 | 103000 | 0.0003 | | 2.8132 | 103500 | 0.0002 | | 2.8268 | 104000 | 0.0002 | | 2.8404 | 104500 | 0.0002 | | 2.8540 | 105000 | 0.0002 | | 2.8675 | 105500 | 0.0001 | | 2.8811 | 106000 | 0.0002 | | 2.8947 | 106500 | 0.0002 | | 2.9083 | 107000 | 0.0002 | | 2.9219 | 107500 | 0.0001 | | 2.9355 | 108000 | 0.0002 | | 2.9491 | 108500 | 0.0003 | | 2.9627 | 109000 | 0.0003 | | 2.9763 | 109500 | 0.0002 | | 2.9899 | 110000 | 0.0002 |

Framework Versions

• Python: 3.10.13
• Sentence Transformers: 3.0.1
• Transformers: 4.45.0.dev0
• PyTorch: 2.2.1
• Accelerate: 0.31.0
• Datasets: 2.17.1
• Tokenizers: 0.19.1

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",
}