reranker-msmarco-v1.1-ettin-encoder-150m-listnet

CrossEncoder based on jhu-clsp/ettin-encoder-150m

This is a Cross Encoder model finetuned from jhu-clsp/ettin-encoder-150m on the ms_marco dataset using the sentence-transformers library. It computes scores for pairs of texts, which can be used for text reranking and semantic search.

Model Details

Model Description

• Model Type: Cross Encoder
• Base model: jhu-clsp/ettin-encoder-150m <!-- at revision 45d08642849e5c5701b162671ac811b7654bfd9f -->
• Maximum Sequence Length: 7999 tokens
• Number of Output Labels: 1 label
• Training Dataset:
  • ms_marco
• Language: en

<!-- - **License:** Unknown -->

Model Sources

• Documentation: Sentence Transformers Documentation
• Documentation: Cross Encoder Documentation
• Repository: Sentence Transformers on GitHub
• Hugging Face: Cross Encoders on Hugging Face

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 CrossEncoder

# Download from the 🤗 Hub
model = CrossEncoder("bansalaman18/reranker-msmarco-v1.1-ettin-encoder-150m-listnet")
# Get scores for pairs of texts
pairs = [
    ['what is the definition of Chirality', 'Two enantiomers of a generic amino acid that is chiral. Chirality /kaɪˈraelɪtiː/ kaɪˈrælɪtiː is a property of symmetry important in several branches of. Science the word chirality is derived from The, greek (χειρ), kheir, hand a familiar chiral. object An object or a system is chiral if it is distinguishable from its mirror image; that is, it cannot be superposed onto it. Conversely, a mirror image of an achiral object, such as a sphere, cannot be distinguished from the object.'],
    ['what is the definition of Chirality', 'The symmetry of a molecule (or any other object) determines whether it is chiral. A molecule is achiral (not chiral) when an improper rotation, that is a combination of a rotation and a reflection in a plane, perpendicular to the axis of rotation, results in the same molecule. For tetrahedral molecules, the molecule is chiral if all four substituents are different. Chirality is important in context of ordered phases as well, for example the addition of a small amount of an optically active molecule to a nematic phase (a phase that has long range orientational order of molecules) transforms that phase to a chiral nematic phase (or cholesteric phase).'],
    ['what is the definition of Chirality', 'Chirality is when an object is present in two forms which are mirror images of each other. The two forms cannot be superimposed. Human hands are chiral. The left hand is a mirror image of the right hand. You cannot turn or move one of your hands to look exactly the same as the other. In chemistry, chirality refers to molecules. If a molecule is chiral it has two forms that are mirror images of each other but otherwise identical. Chemists call these enantiomers or optical isomers.'],
    ['what is the definition of Chirality', 'Chirality /kaɪˈraelɪtiː/ kaɪˈrælɪtiː is a property of symmetry important in several branches of. Science the word chirality is derived from The, greek (χειρ), kheir, hand a familiar chiral. Object an object or a system is chiral if it is distinguishable from its mirror; image that, is it cannot be superposed onto. it An object or a system is chiral if it is distinguishable from its mirror image; that is, it cannot be superposed onto it. Conversely, a mirror image of an achiral object, such as a sphere, cannot be distinguished from the object.'],
    ['what is the definition of Chirality', 'Normally, when a tetrahedral atom has four different substituents it is chiral. However, in rare cases, if two of the ligands differ from each other by being mirror images of each other, the mirror image of the molecule is identical to the original, and the molecule is achiral. Chirality is important in context of ordered phases as well, for example the addition of a small amount of an optically active molecule to a nematic phase (a phase that has long range orientational order of molecules) transforms that phase to a chiral nematic phase (or cholesteric phase).'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'what is the definition of Chirality',
    [
        'Two enantiomers of a generic amino acid that is chiral. Chirality /kaɪˈraelɪtiː/ kaɪˈrælɪtiː is a property of symmetry important in several branches of. Science the word chirality is derived from The, greek (χειρ), kheir, hand a familiar chiral. object An object or a system is chiral if it is distinguishable from its mirror image; that is, it cannot be superposed onto it. Conversely, a mirror image of an achiral object, such as a sphere, cannot be distinguished from the object.',
        'The symmetry of a molecule (or any other object) determines whether it is chiral. A molecule is achiral (not chiral) when an improper rotation, that is a combination of a rotation and a reflection in a plane, perpendicular to the axis of rotation, results in the same molecule. For tetrahedral molecules, the molecule is chiral if all four substituents are different. Chirality is important in context of ordered phases as well, for example the addition of a small amount of an optically active molecule to a nematic phase (a phase that has long range orientational order of molecules) transforms that phase to a chiral nematic phase (or cholesteric phase).',
        'Chirality is when an object is present in two forms which are mirror images of each other. The two forms cannot be superimposed. Human hands are chiral. The left hand is a mirror image of the right hand. You cannot turn or move one of your hands to look exactly the same as the other. In chemistry, chirality refers to molecules. If a molecule is chiral it has two forms that are mirror images of each other but otherwise identical. Chemists call these enantiomers or optical isomers.',
        'Chirality /kaɪˈraelɪtiː/ kaɪˈrælɪtiː is a property of symmetry important in several branches of. Science the word chirality is derived from The, greek (χειρ), kheir, hand a familiar chiral. Object an object or a system is chiral if it is distinguishable from its mirror; image that, is it cannot be superposed onto. it An object or a system is chiral if it is distinguishable from its mirror image; that is, it cannot be superposed onto it. Conversely, a mirror image of an achiral object, such as a sphere, cannot be distinguished from the object.',
        'Normally, when a tetrahedral atom has four different substituents it is chiral. However, in rare cases, if two of the ligands differ from each other by being mirror images of each other, the mirror image of the molecule is identical to the original, and the molecule is achiral. Chirality is important in context of ordered phases as well, for example the addition of a small amount of an optically active molecule to a nematic phase (a phase that has long range orientational order of molecules) transforms that phase to a chiral nematic phase (or cholesteric phase).',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

<!-- ### 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

ms_marco

• Dataset: ms_marco at a47ee7a
• Size: 78,704 training samples
• Columns: <code>query</code>, <code>docs</code>, and <code>labels</code>
• Approximate statistics based on the first 1000 samples:

  	query	docs	labels
  type	string	list	list
  details	<ul><li>min: 11 characters</li><li>mean: 34.52 characters</li><li>max: 112 characters</li></ul>	<ul><li>min: 2 elements</li><li>mean: 6.00 elements</li><li>max: 10 elements</li></ul>	<ul><li>min: 2 elements</li><li>mean: 6.00 elements</li><li>max: 10 elements</li></ul>

• Samples:

  query	docs	labels
  <code>what tissues does the respiratory membrane consist of</code>	<code>['Confidence votes 2.3K. The respiratory membrane consists of the epithelial cells of the alveolus, the endothelial cells of the capillary, and the two fused basement membranes of these layers. ', 'Membranes. Body membranes are thin sheets of tissue that cover the body, line body cavities, and cover organs within the cavities in hollow organs. They can be categorized into epithelial and connective tissue membrane. Synovial membranes are connective tissue membranes that line the cavities of the freely movable joints such as the shoulder, elbow, and knee. Like serous membranes, they line cavities that do not open to the outside.', 'On another level, epithelia and connective tissues combine to form membranes that cover and protect other structures and tissues. Types of membranes: 1 Mucous membranes line cavities that connect with the exterior, including the digestive, respiratory, reproductive, and urinary tracts. 2 The epithelial surfaces are kept moist at all times. 3 The connective ...</code>	<code>[1, 0, 0, 0, 0, ...]</code>
  <code>where is the human stomach located</code>	<code>['All rights reserved. The stomach is a muscular organ located on the left side of the upper abdomen. The stomach receives food from the esophagus. As food reaches the end of the esophagus, it enters the stomach through a muscular valve called the lower esophageal sphincter. The stomach secretes acid and enzymes that digest food. ', 'Stomach. The stomach is a muscular, J-shaped organ in the abdomen. Part of the digestive system, the stomach stores and digests food through gastric juices and a specialized churning action created by folds on its inside. The stomach is located in the upper-left area of the abdomen below the liver and next to the spleen. It’s main function is store and breakdown foods and liquids that we consume before those contents travel to other organs to be further digested. When the stomach is empty, the inside has small folds called rugae.', 'The. Encyclopaedia Encyclopædia, Britannica. Inc the human stomach is Roughly-j shaped and is located in the upper left side ...</code>	<code>[1, 0, 0, 0, 0, ...]</code>
  <code>brick wall cost per linear foot</code>	<code>['The garden is to be enclosed by a stone wall costing $26 per linear foot. The interior wall is to be constructed with brick costing $8 per linear foot. I need to make a quote for a garden wall. The wall is to be one brick thick, 8 metres long and 0.75 meters high above ground level. the ground in the area is good quality top soil to a depth of 500mm and below is solid chalk.', 'Our free calculator uses recent, trusted data to estimate costs for your Concrete Block Wall Cap Installation project. For a basic 25 feet project in zip code 47474, the benchmark cost to Install a Concrete Block Wall Cap ranges between $13.26 - $16.36 per linear foot. To estimate costs for your project: 1. Set Project Zip Code Enter the Zip Code for the location where labor is hired and materials purchased. 2.', 'The cost of an eight foot tall brick fence will be $7.50 a linear foot. This cost is inclusive of material only. It is always best to consult a local building contractor to get a proper estimate of t...</code>	<code>[1, 0, 0, 0, 0, ...]</code>

• Loss: <code>ListNetLoss</code> with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "mini_batch_size": 16
  }
  

Evaluation Dataset

ms_marco

• Dataset: ms_marco at a47ee7a
• Size: 1,000 evaluation samples
• Columns: <code>query</code>, <code>docs</code>, and <code>labels</code>
• Approximate statistics based on the first 1000 samples:

  	query	docs	labels
  type	string	list	list
  details	<ul><li>min: 11 characters</li><li>mean: 33.78 characters</li><li>max: 91 characters</li></ul>	<ul><li>min: 3 elements</li><li>mean: 6.50 elements</li><li>max: 10 elements</li></ul>	<ul><li>min: 3 elements</li><li>mean: 6.50 elements</li><li>max: 10 elements</li></ul>

• Samples:

  query	docs	labels
  <code>what is the definition of Chirality</code>	<code>['Two enantiomers of a generic amino acid that is chiral. Chirality /kaɪˈraelɪtiː/ kaɪˈrælɪtiː is a property of symmetry important in several branches of. Science the word chirality is derived from The, greek (χειρ), kheir, hand a familiar chiral. object An object or a system is chiral if it is distinguishable from its mirror image; that is, it cannot be superposed onto it. Conversely, a mirror image of an achiral object, such as a sphere, cannot be distinguished from the object.', 'The symmetry of a molecule (or any other object) determines whether it is chiral. A molecule is achiral (not chiral) when an improper rotation, that is a combination of a rotation and a reflection in a plane, perpendicular to the axis of rotation, results in the same molecule. For tetrahedral molecules, the molecule is chiral if all four substituents are different. Chirality is important in context of ordered phases as well, for example the addition of a small amount of an optically active molecule to a nem...</code>	<code>[1, 0, 0, 0, 0, ...]</code>
  <code>what is secularism</code>	<code>['Secularism is a code of duty pertaining to this life, founded on considerations purely human, and intended mainly for those who find theology indefinite or inadequate, unreliable or unbelievable. Its essential principles are three: (1) The improvement of this life by material means. In political terms, secularism is a movement towards the separation of religion and government (often termed the separation of church and state).', 'Secularism is about democracy and fairness. In a secular democracy all citizens are equal before the law and parliament. No religious or political affiliation gives advantages or disadvantages and religious believers are citizens with the same rights and obligations as anyone else. Secularism champions human rights above discriminatory religious demands. Secularism seeks to ensure and protect freedom of religious belief and practice for all citizens. Secularism is not about curtailing religious freedoms; it is about ensuring that the freedoms of thought and c...</code>	<code>[1, 0, 0, 0, 0, ...]</code>
  <code>possible permanent side effects of testosterone injections</code>	<code>['Common side effects of Depo-Testosterone include nausea, vomiting, headache, anxiety, depression, skin color changes, increased/decreased sexual interest, oily skin, hair loss, breast swelling in men, numbness or tingly feeling, and acne. Pain, swelling, and redness at the injection site may also occur. This is not a complete list of side effects and others may occur. Tell your doctor about any unusual or bothersome side effect. You may report side effects to FDA at 1-800-FDA-1088. Read the entire detailed patient monograph for Depo-Testosterone (Testosterone Cypionate Injection).', 'What causes side effects. Hormone therapies for prostate cancer cause side effects because they reduce the levels of male hormones in the body. Some side effects are common to all hormone therapies used for prostate cancer. Some effects vary from drug to drug', 'Testosterone replacement therapy is designed to restore your testosterone to normal levels. It is administered through injections, patches, or g...</code>	<code>[1, 0, 0, 0, 0, ...]</code>

• Loss: <code>ListNetLoss</code> with these parameters:

  {
      "activation_fn": "torch.nn.modules.linear.Identity",
      "mini_batch_size": 16
  }
  

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: 5
• seed: 12
• bf16: True
• load_best_model_at_end: True

All Hyperparameters

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

• 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
• torch_empty_cache_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: 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: 12
• data_seed: None
• jit_mode_eval: False
• use_ipex: False
• bf16: True
• 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}
• tp_size: 0
• 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: proportional
• router_mapping: {}
• learning_rate_mapping: {}

</details>

Training Logs

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

• The bold row denotes the saved checkpoint.

</details>

Framework Versions

• Python: 3.11.13
• Sentence Transformers: 5.0.0
• Transformers: 4.51.0
• PyTorch: 2.9.1+cu126
• Accelerate: 1.8.1
• Datasets: 3.6.0
• Tokenizers: 0.21.4-dev.0

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",
}

ListNetLoss

@inproceedings{cao2007learning,
    title={Learning to Rank: From Pairwise Approach to Listwise Approach},
    author={Cao, Zhe and Qin, Tao and Liu, Tie-Yan and Tsai, Ming-Feng and Li, Hang},
    booktitle={Proceedings of the 24th international conference on Machine learning},
    pages={129--136},
    year={2007}
}

<!-- ## 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.* -->