--- base_model: google/electra-large-discriminator datasets: - PiC/phrase_similarity language: - en library_name: sentence-transformers metrics: - cosine_accuracy - cosine_accuracy_threshold - cosine_f1 - cosine_f1_threshold - cosine_precision - cosine_recall - cosine_ap - dot_accuracy - dot_accuracy_threshold - dot_f1 - dot_f1_threshold - dot_precision - dot_recall - dot_ap - manhattan_accuracy - manhattan_accuracy_threshold - manhattan_f1 - manhattan_f1_threshold - manhattan_precision - manhattan_recall - manhattan_ap - euclidean_accuracy - euclidean_accuracy_threshold - euclidean_f1 - euclidean_f1_threshold - euclidean_precision - euclidean_recall - euclidean_ap - max_accuracy - max_accuracy_threshold - max_f1 - max_f1_threshold - max_precision - max_recall - max_ap pipeline_tag: sentence-similarity tags: - sentence-transformers - sentence-similarity - feature-extraction - generated_from_trainer - dataset_size:7004 - loss:SoftmaxLoss widget: - source_sentence: Google SEO expert Matt Cutts had a similar experience, of the eight magazines and newspapers Cutts tried to order, he received zero. sentences: - He dissolved the services of her guards and her court attendants and seized an expansive reach of properties belonging to her. - Google SEO expert Matt Cutts had a comparable occurrence, of the eight magazines and newspapers Cutts tried to order, he received zero. - bill's newest solo play, "all over the map", premiered off broadway in april 2016, produced by all for an individual cinema. - source_sentence: Shula said that Namath "beat our blitz" with his fast release, which let him quickly dump the football off to a receiver. sentences: - Shula said that Namath "beat our blitz" with his quick throw, which let him quickly dump the football off to a receiver. - it elects a single component of parliament (mp) by the first past the post system of election. - Matt Groening said that West was one of the most widely known group to ever come to the studio. - source_sentence: When Angel calls out her name, Cordelia suddenly appears from the opposite side of the room saying, "Yep, that chick's in rough shape. sentences: - The ruined row of text, part of the Florida East Coast Railway, was repaired by 2014 renewing freight train access to the port. - When Angel calls out her name, Cordelia suddenly appears from the opposite side of the room saying, "Yep, that chick's in approximate form. - Chaplin's films introduced a moderated kind of comedy than the typical Keystone farce, and he developed a large fan base. - source_sentence: The following table shows the distances traversed by National Route 11 in each different department, showing cities and towns that it passes by (or near). sentences: - The following table shows the distances traversed by National Route 11 in each separate city authority, showing cities and towns that it passes by (or near). - Similarly, indigenous communities and leaders practice as the main rule of law on local native lands and reserves. - later, sylvan mixed gary numan's albums "replicas" (with numan's previous band tubeway army) and "the quest for instant gratification". - source_sentence: She wants to write about Keima but suffers a major case of writer's block. sentences: - In some countries, new extremist parties on the extreme opposite of left of the political spectrum arose, motivated through issues of immigration, multiculturalism and integration. - specific medical status of movement and the general condition of movement both are conditions under which contradictions can move. - She wants to write about Keima but suffers a huge occurrence of writer's block. model-index: - name: SentenceTransformer based on google/electra-large-discriminator results: - task: type: binary-classification name: Binary Classification dataset: name: quora duplicates dev type: quora-duplicates-dev metrics: - type: cosine_accuracy value: 0.746 name: Cosine Accuracy - type: cosine_accuracy_threshold value: 0.957567036151886 name: Cosine Accuracy Threshold - type: cosine_f1 value: 0.7678410117434508 name: Cosine F1 - type: cosine_f1_threshold value: 0.9061176776885986 name: Cosine F1 Threshold - type: cosine_precision value: 0.700164744645799 name: Cosine Precision - type: cosine_recall value: 0.85 name: Cosine Recall - type: cosine_ap value: 0.8056677266949036 name: Cosine Ap - type: dot_accuracy value: 0.689 name: Dot Accuracy - type: dot_accuracy_threshold value: 256.4343566894531 name: Dot Accuracy Threshold - type: dot_f1 value: 0.7481781376518218 name: Dot F1 - type: dot_f1_threshold value: 252.6839599609375 name: Dot F1 Threshold - type: dot_precision value: 0.6285714285714286 name: Dot Precision - type: dot_recall value: 0.924 name: Dot Recall - type: dot_ap value: 0.6031219278518416 name: Dot Ap - type: manhattan_accuracy value: 0.746 name: Manhattan Accuracy - type: manhattan_accuracy_threshold value: 103.93816375732422 name: Manhattan Accuracy Threshold - type: manhattan_f1 value: 0.767172167707404 name: Manhattan F1 - type: manhattan_f1_threshold value: 200.26837158203125 name: Manhattan F1 Threshold - type: manhattan_precision value: 0.6924315619967794 name: Manhattan Precision - type: manhattan_recall value: 0.86 name: Manhattan Recall - type: manhattan_ap value: 0.8064055162262082 name: Manhattan Ap - type: euclidean_accuracy value: 0.744 name: Euclidean Accuracy - type: euclidean_accuracy_threshold value: 5.6128997802734375 name: Euclidean Accuracy Threshold - type: euclidean_f1 value: 0.7639639639639639 name: Euclidean F1 - type: euclidean_f1_threshold value: 8.71990966796875 name: Euclidean F1 Threshold - type: euclidean_precision value: 0.6950819672131148 name: Euclidean Precision - type: euclidean_recall value: 0.848 name: Euclidean Recall - type: euclidean_ap value: 0.8072457371353838 name: Euclidean Ap - type: max_accuracy value: 0.746 name: Max Accuracy - type: max_accuracy_threshold value: 256.4343566894531 name: Max Accuracy Threshold - type: max_f1 value: 0.7678410117434508 name: Max F1 - type: max_f1_threshold value: 252.6839599609375 name: Max F1 Threshold - type: max_precision value: 0.700164744645799 name: Max Precision - type: max_recall value: 0.924 name: Max Recall - type: max_ap value: 0.8072457371353838 name: Max Ap --- # SentenceTransformer based on google/electra-large-discriminator This is a [sentence-transformers](https://www.SBERT.net) model finetuned from [google/electra-large-discriminator](https://huggingface.co/google/electra-large-discriminator) on the [PiC/phrase_similarity](https://huggingface.co/datasets/PiC/phrase_similarity) 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:** [google/electra-large-discriminator](https://huggingface.co/google/electra-large-discriminator) - **Maximum Sequence Length:** 512 tokens - **Output Dimensionality:** 1024 tokens - **Similarity Function:** Cosine Similarity - **Training Dataset:** - [PiC/phrase_similarity](https://huggingface.co/datasets/PiC/phrase_similarity) - **Language:** en ### Model Sources - **Documentation:** [Sentence Transformers Documentation](https://sbert.net) - **Repository:** [Sentence Transformers on GitHub](https://github.com/UKPLab/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': 512, 'do_lower_case': False}) with Transformer model: ElectraModel (1): Pooling({'word_embedding_dimension': 1024, '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: ```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("Deehan1866/finetuned-valloss-electra-large-discriminator") # Run inference sentences = [ "She wants to write about Keima but suffers a major case of writer's block.", "She wants to write about Keima but suffers a huge occurrence of writer's block.", 'specific medical status of movement and the general condition of movement both are conditions under which contradictions can move.', ] embeddings = model.encode(sentences) print(embeddings.shape) # [3, 1024] # Get the similarity scores for the embeddings similarities = model.similarity(embeddings, embeddings) print(similarities.shape) # [3, 3] ``` ## Evaluation ### Metrics #### Binary Classification * Dataset: `quora-duplicates-dev` * Evaluated with [BinaryClassificationEvaluator](https://sbert.net/docs/package_reference/sentence_transformer/evaluation.html#sentence_transformers.evaluation.BinaryClassificationEvaluator) | Metric | Value | |:-----------------------------|:-----------| | cosine_accuracy | 0.746 | | cosine_accuracy_threshold | 0.9576 | | cosine_f1 | 0.7678 | | cosine_f1_threshold | 0.9061 | | cosine_precision | 0.7002 | | cosine_recall | 0.85 | | cosine_ap | 0.8057 | | dot_accuracy | 0.689 | | dot_accuracy_threshold | 256.4344 | | dot_f1 | 0.7482 | | dot_f1_threshold | 252.684 | | dot_precision | 0.6286 | | dot_recall | 0.924 | | dot_ap | 0.6031 | | manhattan_accuracy | 0.746 | | manhattan_accuracy_threshold | 103.9382 | | manhattan_f1 | 0.7672 | | manhattan_f1_threshold | 200.2684 | | manhattan_precision | 0.6924 | | manhattan_recall | 0.86 | | manhattan_ap | 0.8064 | | euclidean_accuracy | 0.744 | | euclidean_accuracy_threshold | 5.6129 | | euclidean_f1 | 0.764 | | euclidean_f1_threshold | 8.7199 | | euclidean_precision | 0.6951 | | euclidean_recall | 0.848 | | euclidean_ap | 0.8072 | | max_accuracy | 0.746 | | max_accuracy_threshold | 256.4344 | | max_f1 | 0.7678 | | max_f1_threshold | 252.684 | | max_precision | 0.7002 | | max_recall | 0.924 | | **max_ap** | **0.8072** | ## Training Details ### Training Dataset #### PiC/phrase_similarity * Dataset: [PiC/phrase_similarity](https://huggingface.co/datasets/PiC/phrase_similarity) at [fc67ce7](https://huggingface.co/datasets/PiC/phrase_similarity/tree/fc67ce7c1e69e360e42dc6f31ddf97bb32f1923d) * Size: 7,004 training samples * Columns: sentence1, sentence2, and label * Approximate statistics based on the first 1000 samples: | | sentence1 | sentence2 | label | |:--------|:-----------------------------------------------------------------------------------|:-----------------------------------------------------------------------------------|:------------------------------------------------| | type | string | string | int | | details | | | | * Samples: | sentence1 | sentence2 | label | |:------------------------------------------------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------|:---------------| | newly formed camp is released from the membrane and diffuses across the intracellular space where it serves to activate pka. | recently made encampment is released from the membrane and diffuses across the intracellular space where it serves to activate pka. | 0 | | According to one data, in 1910, on others – in 1915, the mansion became Natalya Dmitriyevna Shchuchkina's property. | According to a particular statistic, in 1910, on others – in 1915, the mansion became Natalya Dmitriyevna Shchuchkina's property. | 1 | | Note that Fact 1 does not assume any particular structure on the set formula_65. | Note that Fact 1 does not assume any specific edifice on the set formula_65. | 0 | * Loss: [SoftmaxLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#softmaxloss) ### Evaluation Dataset #### PiC/phrase_similarity * Dataset: [PiC/phrase_similarity](https://huggingface.co/datasets/PiC/phrase_similarity) at [fc67ce7](https://huggingface.co/datasets/PiC/phrase_similarity/tree/fc67ce7c1e69e360e42dc6f31ddf97bb32f1923d) * Size: 1,000 evaluation samples * Columns: sentence1, sentence2, and label * Approximate statistics based on the first 1000 samples: | | sentence1 | sentence2 | label | |:--------|:----------------------------------------------------------------------------------|:----------------------------------------------------------------------------------|:------------------------------------------------| | type | string | string | int | | details | | | | * Samples: | sentence1 | sentence2 | label | |:----------------------------------------------------------------------------------------------------------------------------|:------------------------------------------------------------------------------------------------------------------------------|:---------------| | after theo's apparent death, she decides to leave first colony and ends up traveling with the apostles. | after theo's apparent death, she decides to leave original settlement and ends up traveling with the apostles. | 0 | | The guard assigned to Vivian leaves her to prevent the robbery, allowing her to connect to the bank's network. | The guard assigned to Vivian leaves her to prevent the robbery, allowing her to connect to the bank's locations. | 0 | | Two days later Louis XVI banished Necker by a "lettre de cachet" for his very public exchange of pamphlets. | Two days later Louis XVI banished Necker by a "lettre de cachet" for his very free forum of pamphlets. | 0 | * Loss: [SoftmaxLoss](https://sbert.net/docs/package_reference/sentence_transformer/losses.html#softmaxloss) ### Training Hyperparameters #### Non-Default Hyperparameters - `eval_strategy`: steps - `per_device_train_batch_size`: 16 - `per_device_eval_batch_size`: 16 - `learning_rate`: 2e-05 - `num_train_epochs`: 100 - `warmup_ratio`: 0.1 - `load_best_model_at_end`: True #### All Hyperparameters
Click to expand - `overwrite_output_dir`: False - `do_predict`: False - `eval_strategy`: steps - `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 - `learning_rate`: 2e-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`: 100 - `max_steps`: -1 - `lr_scheduler_type`: linear - `lr_scheduler_kwargs`: {} - `warmup_ratio`: 0.1 - `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`: True - `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 - `batch_sampler`: batch_sampler - `multi_dataset_batch_sampler`: proportional
### Training Logs | Epoch | Step | Training Loss | loss | quora-duplicates-dev_max_ap | |:----------:|:-------:|:-------------:|:----------:|:---------------------------:| | 0 | 0 | - | - | 0.6721 | | 0.2283 | 100 | - | 0.6942 | 0.6722 | | 0.4566 | 200 | - | 0.6932 | 0.6727 | | 0.6849 | 300 | - | 0.6889 | 0.6730 | | 0.9132 | 400 | - | 0.6813 | 0.6933 | | 1.1416 | 500 | 0.6891 | 0.6562 | 0.7557 | | 1.3699 | 600 | - | 0.5780 | 0.7861 | | 1.5982 | 700 | - | 0.5417 | 0.7894 | | **1.8265** | **800** | **-** | **0.5238** | **0.8072** | | 2.0548 | 900 | - | 0.5261 | 0.8079 | | 2.2831 | 1000 | 0.5426 | 0.5418 | 0.8144 | | 2.5114 | 1100 | - | 0.5463 | 0.8161 | | 2.7397 | 1200 | - | 0.5737 | 0.7916 | | 2.9680 | 1300 | - | 0.5558 | 0.8038 | | 3.1963 | 1400 | - | 0.5878 | 0.8175 | | 3.4247 | 1500 | 0.4201 | 0.6230 | 0.8040 | | 3.6530 | 1600 | - | 0.6078 | 0.8064 | | 3.8813 | 1700 | - | 0.6065 | 0.8042 | | 4.1096 | 1800 | - | 0.7494 | 0.8072 | * The bold row denotes the saved checkpoint. ### Framework Versions - Python: 3.10.10 - Sentence Transformers: 3.0.1 - Transformers: 4.42.3 - PyTorch: 2.2.1+cu121 - Accelerate: 0.32.1 - Datasets: 2.20.0 - Tokenizers: 0.19.1 ## Citation ### BibTeX #### Sentence Transformers and SoftmaxLoss ```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", } ```