SentenceTransformer(
(0): Transformer({'transformer_task': 'feature-extraction', 'modality_config': {'text': {'method': 'forward', 'method_output_name': 'last_hidden_state'}}, 'module_output_name': 'token_embeddings', 'architecture': 'BertModel'})
(1): Pooling({'embedding_dimension': 384, 'pooling_mode': 'mean', 'include_prompt': True})
(2): Normalize({})
)pip install -U sentence-transformers1from sentence_transformers import SentenceTransformer
2
3# Download from the 🤗 Hub
4model = SentenceTransformer("MeherMs/legal-contract-similarity")
5# Run inference
6sentences = [
7 'Each party agrees to keep confidential all proprietary information disclosed by the other party and to use such information solely for the purposes of this agreement.',
8 'All information shared between parties may be freely distributed to third parties without restriction.',
9 'The client shall own all intellectual property rights to work created by the contractor under this engagement.',
10]
11embeddings = model.encode(sentences)
12print(embeddings.shape)
13# [3, 384]
14
15# Get the similarity scores for the embeddings
16similarities = model.similarity(embeddings, embeddings)
17print(similarities)
18# tensor([[1.0000, 0.2213, 0.2642],
19# [0.2213, 1.0000, 0.1867],
20# [0.2642, 0.1867, 1.0000]])legal-similarity-v1EmbeddingSimilarityEvaluator| Metric | Value |
|---|---|
| pearson_cosine | 1.0 |
| spearman_cosine | 1.0 |
sentence_0, sentence_1, and label| sentence_0 | sentence_1 | label | |
|---|---|---|---|
| type | string | string | float |
| modality | text | text | |
| details |
|
|
|
| sentence_0 | sentence_1 | label |
|---|---|---|
Each party agrees to keep confidential all proprietary information disclosed by the other party and to use such information solely for the purposes of this agreement. | All information shared between parties may be freely distributed to third parties without restriction. | 0.3 |
The company assumes full responsibility for all damages caused by negligence or breach of contract. | All information shared between parties may be freely distributed to third parties without restriction. | 0.1 |
Each party agrees to keep confidential all proprietary information disclosed by the other party and to use such information solely for the purposes of this agreement. | The receiving party shall maintain confidentiality of all proprietary information shared by the disclosing party and shall use it only as permitted under this agreement. | 0.8 |
CosineSimilarityLoss with these parameters:
1{
2 "loss_fct": "torch.nn.modules.loss.MSELoss",
3 "cos_score_transformation": "torch.nn.modules.linear.Identity"
4}num_train_epochs: 20disable_tqdm: Truemulti_dataset_batch_sampler: round_robinper_device_train_batch_size: 8num_train_epochs: 20max_steps: -1learning_rate: 5e-05lr_scheduler_type: linearlr_scheduler_kwargs: Nonewarmup_steps: 0optim: adamw_torch_fusedoptim_args: Noneweight_decay: 0.0adam_beta1: 0.9adam_beta2: 0.999adam_epsilon: 1e-08optim_target_modules: Nonegradient_accumulation_steps: 1average_tokens_across_devices: Truemax_grad_norm: 1label_smoothing_factor: 0.0bf16: Falsefp16: Falsebf16_full_eval: Falsefp16_full_eval: Falsetf32: Nonegradient_checkpointing: Falsegradient_checkpointing_kwargs: Nonetorch_compile: Falsetorch_compile_backend: Nonetorch_compile_mode: Noneuse_liger_kernel: Falseliger_kernel_config: Noneuse_cache: Falseneftune_noise_alpha: Nonetorch_empty_cache_steps: Noneauto_find_batch_size: Falselog_on_each_node: Truelogging_nan_inf_filter: Trueinclude_num_input_tokens_seen: nolog_level: passivelog_level_replica: warningdisable_tqdm: Trueproject: huggingfacetrackio_space_id: Nonetrackio_bucket_id: Nonetrackio_static_space_id: Noneper_device_eval_batch_size: 8prediction_loss_only: Trueeval_on_start: Falseeval_do_concat_batches: Trueeval_use_gather_object: Falseeval_accumulation_steps: Noneinclude_for_metrics: []batch_eval_metrics: Falsesave_only_model: Falsesave_on_each_node: Falseenable_jit_checkpoint: Falsepush_to_hub: Falsehub_private_repo: Nonehub_model_id: Nonehub_strategy: every_savehub_always_push: Falsehub_revision: Noneload_best_model_at_end: Falseignore_data_skip: Falserestore_callback_states_from_checkpoint: Falsefull_determinism: Falseseed: 42data_seed: Noneuse_cpu: Falseaccelerator_config: {'split_batches': False, 'dispatch_batches': None, 'even_batches': True, 'use_seedable_sampler': True, 'non_blocking': False, 'gradient_accumulation_kwargs': None}parallelism_config: Nonedataloader_drop_last: Falsedataloader_num_workers: 0dataloader_pin_memory: Truedataloader_persistent_workers: Falsedataloader_prefetch_factor: Noneremove_unused_columns: Truelabel_names: Nonetrain_sampling_strategy: randomlength_column_name: lengthddp_find_unused_parameters: Noneddp_bucket_cap_mb: Noneddp_broadcast_buffers: Falseddp_static_graph: Noneddp_backend: Noneddp_timeout: 1800fsdp: Nonefsdp_config: Nonedeepspeed: Nonedebug: []skip_memory_metrics: Truedo_predict: Falseresume_from_checkpoint: Nonewarmup_ratio: Nonelocal_rank: -1prompts: Nonebatch_sampler: batch_samplermulti_dataset_batch_sampler: round_robinrouter_mapping: {}learning_rate_mapping: {}| Epoch | Step | legal-similarity-v1_spearman_cosine |
|---|---|---|
| 1.0 | 2 | 1.0000 |
| 2.0 | 4 | 1.0000 |
| 2.5 | 5 | 1.0000 |
| 3.0 | 6 | 1.0000 |
| 4.0 | 8 | 1.0000 |
| 5.0 | 10 | 1.0000 |
| 6.0 | 12 | 1.0000 |
| 7.0 | 14 | 1.0000 |
| 7.5 | 15 | 1.0000 |
| 8.0 | 16 | 1.0000 |
| 9.0 | 18 | 1.0000 |
| 10.0 | 20 | 1.0000 |
| 11.0 | 22 | 1.0000 |
| 12.0 | 24 | 1.0000 |
| 12.5 | 25 | 1.0000 |
| 13.0 | 26 | 1.0000 |
| 14.0 | 28 | 1.0000 |
| 15.0 | 30 | 1.0000 |
| 16.0 | 32 | 1.0000 |
| 17.0 | 34 | 1.0000 |
| 17.5 | 35 | 1.0000 |
| 18.0 | 36 | 1.0000 |
| 19.0 | 38 | 1.0000 |
| 20.0 | 40 | 1.0000 |
1@inproceedings{reimers-2019-sentence-bert,
2 title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
3 author = "Reimers, Nils and Gurevych, Iryna",
4 booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
5 month = "11",
6 year = "2019",
7 publisher = "Association for Computational Linguistics",
8 url = "https://arxiv.org/abs/1908.10084",
9}