DeBERTa-ST-AllLayers-v3.1

SentenceTransformer based on bobox/DeBERTa-ST-AllLayers-v3-checkpoints-tmp

This is a sentence-transformers model finetuned from bobox/DeBERTa-ST-AllLayers-v3-checkpoints-tmp on the nli-pairs, sts-label, vitaminc-pairs, qnli-contrastive, scitail-pairs-qa, scitail-pairs-pos, xsum-pairs, compression-pairs, compression-pairs2, compression-pairs3, sciq_pairs, qasc_pairs, openbookqa_pairs, msmarco_pairs, msmarco_pairs2, nq_pairs, nq_pairs2, trivia_pairs, quora_pairs, gooaq_pairs, gooaq_pairs2 and mrpc_pairs datasets. 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
• Base model: bobox/DeBERTa-ST-AllLayers-v3-checkpoints-tmp <!-- at revision 4859fef4e21d101c2d445bcd33db1e3308f35dc4 -->
• Maximum Sequence Length: 512 tokens
• Output Dimensionality: 768 tokens
• Similarity Function: Cosine Similarity
• Training Datasets:
  • nli-pairs
  • sts-label
  • vitaminc-pairs
  • qnli-contrastive
  • scitail-pairs-qa
  • scitail-pairs-pos
  • xsum-pairs
  • compression-pairs
  • compression-pairs2
  • compression-pairs3
  • sciq_pairs
  • qasc_pairs
  • openbookqa_pairs
  • msmarco_pairs
  • msmarco_pairs2
  • nq_pairs
  • nq_pairs2
  • trivia_pairs
  • quora_pairs
  • gooaq_pairs
  • gooaq_pairs2
  • mrpc_pairs
• Language: en

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

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: DebertaV2Model 
  (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("bobox/DeBERTa-ST-AllLayers-v3.1")
# Run inference
sentences = [
    'Hydrogen is in gas form at room temperature.',
    'In what form of matter is hydrogen at room temperature?',
    'How hot is it on the surface of the Sun?',
]
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]

<!-- ### 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.* -->

Evaluation

Metrics

Semantic Similarity

• Dataset: sts-test
• Evaluated with <code>EmbeddingSimilarityEvaluator</code>

<!-- ## 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 Datasets

nli-pairs

• Dataset: nli-pairs at d482672
• Size: 18,000 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: 5 tokens</li><li>mean: 16.62 tokens</li><li>max: 62 tokens</li></ul>	<ul><li>min: 4 tokens</li><li>mean: 9.46 tokens</li><li>max: 29 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>A person on a horse jumps over a broken down airplane.</code>	<code>A person is outdoors, on a horse.</code>
  <code>Children smiling and waving at camera</code>	<code>There are children present</code>
  <code>A boy is jumping on skateboard in the middle of a red bridge.</code>	<code>The boy does a skateboarding trick.</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

sts-label

• Dataset: sts-label at ab7a5ac
• Size: 5,749 training samples
• Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>score</code>
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	score
  type	string	string	float
  details	<ul><li>min: 6 tokens</li><li>mean: 9.81 tokens</li><li>max: 27 tokens</li></ul>	<ul><li>min: 5 tokens</li><li>mean: 9.74 tokens</li><li>max: 25 tokens</li></ul>	<ul><li>min: 0.0</li><li>mean: 0.54</li><li>max: 1.0</li></ul>

• Samples:

  sentence1	sentence2	score
  <code>A plane is taking off.</code>	<code>An air plane is taking off.</code>	<code>1.0</code>
  <code>A man is playing a large flute.</code>	<code>A man is playing a flute.</code>	<code>0.76</code>
  <code>A man is spreading shreded cheese on a pizza.</code>	<code>A man is spreading shredded cheese on an uncooked pizza.</code>	<code>0.76</code>

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

  {
      "scale": 20.0,
      "similarity_fct": "pairwise_cos_sim"
  }
  

vitaminc-pairs

• Dataset: vitaminc-pairs at be6febb
• Size: 18,000 training samples
• Columns: <code>claim</code> and <code>evidence</code>
• Approximate statistics based on the first 1000 samples:

  	claim	evidence
  type	string	string
  details	<ul><li>min: 6 tokens</li><li>mean: 16.71 tokens</li><li>max: 54 tokens</li></ul>	<ul><li>min: 8 tokens</li><li>mean: 37.14 tokens</li><li>max: 260 tokens</li></ul>

• Samples:

  claim	evidence
  <code>Manchester had a population of more than 540,000 in 2017 and was the 5th most populous English district .</code>	<code>Manchester ( ) is a major city and metropolitan borough in Greater Manchester , England , with a population of 545,500 as of 2017 ( 5th most populous English district ) .</code>
  <code>Manchester had a population of less than 540,000 in 2018 and was the 4th most populous English district .</code>	<code>Manchester ( ) is a major city and metropolitan borough in Greater Manchester , England , with a population of 534,982 as of 2018 ( 4th most populous English district ) .</code>
  <code>Traditional Chinese medicine is founded on more than 4000 years of ancient Chinese medical science and practice .</code>	<code>Traditional Chinese medicine ( TCM ; ) is an ancient system of medical diagnosis and treatment of illnesses with a holistic focus on disease prevention through diet , healthy lifestyle changes , exercise and is built on a patient centered clinically oriented foundation of more than 6,500 years of ancient Chinese medical science and practice that includes various forms of herbal medicine , acupuncture , massage ( tui na ) , exercise ( qigong ) , and dietary therapy , but recently also influenced by modern Western medicine .</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

qnli-contrastive

• Dataset: qnli-contrastive at bcdcba7
• Size: 17,000 training samples
• Columns: <code>sentence1</code>, <code>sentence2</code>, and <code>label</code>
• Approximate statistics based on the first 1000 samples:

  	sentence1	sentence2	label
  type	string	string	int
  details	<ul><li>min: 6 tokens</li><li>mean: 13.7 tokens</li><li>max: 27 tokens</li></ul>	<ul><li>min: 6 tokens</li><li>mean: 35.42 tokens</li><li>max: 499 tokens</li></ul>	<ul><li>0: 100.00%</li></ul>

• Samples:

  sentence1	sentence2	label
  <code>What kind of sound did Kanye abandon a rap and hiphop one for with his fourth album?</code>	<code>West's fourth studio album, 808s & Heartbreak (2008), marked an even more radical departure from his previous releases, largely abandoning rap and hip hop stylings in favor of a stark electropop sound composed of virtual synthesis, the Roland TR-808 drum machine, and explicitly auto-tuned vocal tracks.</code>	<code>0</code>
  <code>When did hostilities in the Persian Gulf War begin?</code>	<code>The lead up to the war began with the Iraqi invasion of Kuwait in August 1990 which was met with immediate economic sanctions by the United Nations against Iraq.</code>	<code>0</code>
  <code>What were two of Marvel's comic heroes in fantasy, swords and magic settings?</code>	<code>Once again, Marvel attempted to diversify, and with the updating of the Comics Code achieved moderate to strong success with titles themed to horror (The Tomb of Dracula), martial arts, (Shang-Chi: Master of Kung Fu), sword-and-sorcery (Conan the Barbarian, Red Sonja), satire (Howard the Duck) and science fiction (2001:</code>	<code>0</code>

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

  {
      "loss": "OnlineContrastiveLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 0.75,
      "prior_layers_weight": 1,
      "kl_div_weight": 0.9,
      "kl_temperature": 0.75
  }
  

scitail-pairs-qa

• Dataset: scitail-pairs-qa at 0cc4353
• Size: 14,312 training samples
• Columns: <code>sentence2</code> and <code>sentence1</code>
• Approximate statistics based on the first 1000 samples:

  	sentence2	sentence1
  type	string	string
  details	<ul><li>min: 7 tokens</li><li>mean: 15.92 tokens</li><li>max: 41 tokens</li></ul>	<ul><li>min: 6 tokens</li><li>mean: 15.03 tokens</li><li>max: 34 tokens</li></ul>

• Samples:

  sentence2	sentence1
  <code>Thermal energy constitutes the total kinetic energy of all the atoms that make up an object.</code>	<code>What kind of energy constitutes the total kinetic energy of all the atoms that make up an object?</code>
  <code>Overharvesting is a serious threat particularly to aquatic species.</code>	<code>Overharvesting is a serious threat particularly to which species?</code>
  <code>Cellulose is created by the polymerization of glucose.</code>	<code>What is created by the polymerization of glucose?</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

scitail-pairs-pos

• Dataset: scitail-pairs-pos at 0cc4353
• Size: 8,600 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: 5 tokens</li><li>mean: 23.59 tokens</li><li>max: 60 tokens</li></ul>	<ul><li>min: 7 tokens</li><li>mean: 15.53 tokens</li><li>max: 41 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>Most of the ozone in the Earth's atmosphere lies in the stratosphere, the layer above the troposphere.</code>	<code>The stratosphere is the layer above the troposphere.</code>
  <code>Exocytosis, fertilization of an egg by sperm and transport of waste products to the lysozome are a few of the many eukaryotic processes that rely on some form of fusion.</code>	<code>The cell expels waste and other particles through a process called exocytosis.</code>
  <code>Mercury, the smallest planet in the Solar System, has the most eccentric orbit.</code>	<code>Mercury is the smallest planet in our solar system.</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

xsum-pairs

• Dataset: xsum-pairs at 788ddaf
• Size: 4,000 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: 2 tokens</li><li>mean: 197.9 tokens</li><li>max: 400 tokens</li></ul>	<ul><li>min: 10 tokens</li><li>mean: 25.48 tokens</li><li>max: 65 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>Mitchell Starc took 4-36 after captain Steve Smith hit 165 as Australia declared their first innings on 624-8.<br>Australia had 68 overs to force a win. They reduced Pakistan to 91-5 at tea before sealing victory in 53.2 overs to take an unassailable 2-0 series lead.<br>Pakistan captain Misbah-ul-Haq was fined 40% of his match fee and the rest of his team 20% for a slow over rate.<br>"It finished a lot better than it started," Starc said. "It's fantastic for us to get that result. The belief was there and it was a fantastic way to finish.<br>"We knew we only had two sessions to get those 10 wickets and together as a bowling unit we've done really well."<br>The third Test starts in Sydney on 3 January but Misbah has not committed to playing in the match after a poor series so far.<br>The 42-year-old, who was dismissed for a two-ball duck after managing just 11 in the first innings, only scored nine runs in total in the first test in Brisbane.<br>"I haven't decided about that [Sydney] but let's see," he said. "[If I'm not contributing] there's no point in hanging around."</code>	<code>Australia bowled Pakistan out for 163 in Melbourne to win the second Test by an innings and 18 runs on day five.</code>
  <code>The Competition and Markets Authority (CMA) had ordered all insurance companies to split out the extra charges for the additional protection.<br>But the Co-op was the only firm which missed a deadline to do so, in August last year.<br>As a result around 120,000 customers received quotations that were unclear.<br>From 1 February, the Co-op will provide two separate quotations - one with no claims bonus protection, and one without.<br>"It is very disappointing that a major company such as Co-op Insurance has taken so long to provide this vital information to its customers," said Adam Land, senior director of remedies, business and financial analysis at the CMA.<br>"Before the order came into force, the price and benefits of NCB [no claims bonus] protection were often unclear to drivers.<br>"We expect the Co-op to fully comply with the terms of our directions immediately, so that motorists can search more easily for the best deal for them, and decide whether or not they want this optional cover."<br>The Co-op said most of its quotations do now provide separate details of no claims bonus charges.<br>"For 90% of our new business customers we are already fully compliant with this order," a spokesperson said.<br>"We are part way through a major transformation programme, which when complete will allow us to be fully compliant and enable us to provide best in class service to our members."</code>	<code>The Co-op has been ordered to provide clearer insurance quotations, after it failed to tell motorists about separate charges for no claims bonuses.</code>
  <code>The 21-year-old spent last season with the League Two club, scoring six times Argyle reached the play-off final.<br>Tanner was part of the Reading squad which defeated Derek Adams' side 2-0 in the EFL Cup first round on 9 August.<br>Tanner, who signed a two-year contract extension with Reading in January 2015, is eligible for Saturday's home fixture against Mansfield Town.<br>Find all the latest football transfers on our dedicated page.</code>	<code>Plymouth Argyle have re-signed Reading midfielder Craig Tanner on a loan deal until January.</code>

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

  {
      "loss": "MultipleNegativesSymmetricRankingLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 0.75,
      "prior_layers_weight": 1,
      "kl_div_weight": 0.9,
      "kl_temperature": 0.75
  }
  

compression-pairs

• Dataset: compression-pairs at 605bc91
• Size: 10,125 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: 9 tokens</li><li>mean: 32.07 tokens</li><li>max: 512 tokens</li></ul>	<ul><li>min: 5 tokens</li><li>mean: 10.27 tokens</li><li>max: 25 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>Democrat State Senate President John Cullerton will endorse Democrat State Treasurer Alexi Giannoulias for US Senate today.</code>	<code>Cullerton to endorse Giannoulias</code>
  <code>Barbara Walters hospitalised after fall Updated: 18:47, Monday January 21, 2013 Veteran ABC newswoman Barbara Walters has been hospitalised after falling at an inauguration party at the residence of Britain's ambassador to the United States.</code>	<code>Barbara Walters hospitalised after fall</code>
  <code>Girl Next-Door star and Hugh Hefner's ex Bridget Marquardt recently moved out of his Playboy mansion.</code>	<code>Bridget Marquardt moves out of Playboy mansion</code>

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

  {
      "loss": "MultipleNegativesSymmetricRankingLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 0.75,
      "prior_layers_weight": 1,
      "kl_div_weight": 0.9,
      "kl_temperature": 0.75
  }
  

compression-pairs2

• Dataset: compression-pairs2 at 605bc91
• Size: 4,701 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: 11 tokens</li><li>mean: 32.18 tokens</li><li>max: 232 tokens</li></ul>	<ul><li>min: 5 tokens</li><li>mean: 10.24 tokens</li><li>max: 29 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>The actress, Natasha Richardson, has died in hospital in New York after suffering a serious head injury in a skiing accident in Canada earlier this week.</code>	<code>Actress Natasha Richardson dies</code>
  <code>'WHISPERING' Ted Lowe - the most recognisable voice in the history of snooker broadcasting - died four hours before the balls were broken in the final of the sport's most important championship</code>	<code>'Whispering' Ted Lowe dies</code>
  <code>Nairobi - Somalia's Shabaab Islamists said on Thursday they have executed a French agent they have held since 2009, as France said the hostage was likely killed several days ago in a failed rescue attempt.</code>	<code>Shabaab say they have executed French hostage</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": 3,
      "last_layer_weight": 0.25,
      "prior_layers_weight": 2,
      "kl_div_weight": 0.75,
      "kl_temperature": 0.75
  }
  

compression-pairs3

• Dataset: compression-pairs3 at 605bc91
• Size: 4,700 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: 12 tokens</li><li>mean: 31.18 tokens</li><li>max: 99 tokens</li></ul>	<ul><li>min: 5 tokens</li><li>mean: 10.02 tokens</li><li>max: 23 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>Pakistan State Oil has started supply of furnace oil to Karachi Electric Supply Company on discounted rates.</code>	<code>Pso starts supply of oil on discounted rates:</code>
  <code>European stocks were little changed as the region's finance chiefs meet today to work on a new strategy to contain the sovereign-debt crisis.</code>	<code>European stocks little changed as finance chiefs meet;</code>
  <code>The body of a missing boater has been found in Lake Okeechobee in southwest Florida.</code>	<code>Body of missing boater found in Lake Okeechobee</code>

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

  {
      "loss": "MultipleNegativesRankingLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 0.01,
      "prior_layers_weight": 10,
      "kl_div_weight": 3,
      "kl_temperature": 0.25
  }
  

sciq_pairs

• Dataset: sciq_pairs at 2c94ad3
• Size: 11,153 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: 7 tokens</li><li>mean: 17.12 tokens</li><li>max: 61 tokens</li></ul>	<ul><li>min: 2 tokens</li><li>mean: 86.03 tokens</li><li>max: 512 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>Temperature and what other environmental factor are important in the activity of an enzyme?</code>	<code></code>
  <code>The sum of the superscripts in an electron configuration is equal to the number of electrons in that atom, which is in turn equal to what number?</code>	<code>Electron configuration notation eliminates the boxes and arrows of orbital filling diagrams. Each occupied sublevel designation is written followed by a superscript that is the number of electrons in that sublevel. For example, the hydrogen configuration is 1 s 1 , while the helium configuration is 1 s 2 . Multiple occupied sublevels are written one after another. The electron configuration of lithium is 1 s 2 2 s 1 . The sum of the superscripts in an electron configuration is equal to the number of electrons in that atom, which is in turn equal to its atomic number.</code>
  <code>What is the most common type of brain injury?</code>	<code>The most common type of brain injury is a concussion. This is a bruise on the surface of the brain. It may cause temporary symptoms such as headache and confusion. Most concussions heal on their own in a few days or weeks. However, repeated concussions can lead to permanent changes in the brain. More serious brain injuries also often cause permanent brain damage.</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

qasc_pairs

• Dataset: qasc_pairs at a34ba20
• Size: 7,767 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: 5 tokens</li><li>mean: 11.52 tokens</li><li>max: 29 tokens</li></ul>	<ul><li>min: 15 tokens</li><li>mean: 34.76 tokens</li><li>max: 75 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>How are organisms able to grow and repair their cells?</code>	<code>Cell division is how organisms grow and repair themselves.. Mitosis is cell division.. Mitosis is how organisms grow and repair themselves.</code>
  <code>What cannot absorb light energy?</code>	<code>chlorophyll is used for absorbing light energy by plants. Fungi have no chlorophyll.. Fungi cannot absorb light energy.</code>
  <code>what effects the colligative properties of solids</code>	<code>adding salt to a solid decreases the freezing point of that solid. Freezing point depression is a colligative property.. salts effect the colligative properties of solids</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

openbookqa_pairs

• Dataset: openbookqa_pairs
• Size: 4,505 training samples
• Columns: <code>question</code> and <code>fact</code>
• Approximate statistics based on the first 1000 samples:

  	question	fact
  type	string	string
  details	<ul><li>min: 3 tokens</li><li>mean: 13.81 tokens</li><li>max: 78 tokens</li></ul>	<ul><li>min: 4 tokens</li><li>mean: 11.49 tokens</li><li>max: 30 tokens</li></ul>

• Samples:

  question	fact
  <code>What is animal competition?</code>	<code>if two animals eat the same prey then those animals compete for that pey</code>
  <code>If you wanted to make a metal bed frame, where would you start?</code>	<code>alloys are made of two or more metals</code>
  <code>Places lacking warmth have few what</code>	<code>cold environments contain few organisms</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

msmarco_pairs

• Dataset: msmarco_pairs at 28ff31e
• Size: 10,314 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: 4 tokens</li><li>mean: 8.7 tokens</li><li>max: 41 tokens</li></ul>	<ul><li>min: 17 tokens</li><li>mean: 77.31 tokens</li><li>max: 258 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>otc medication meaning</code>	<code>Over-the-counter (OTC) medicines are drugs you can buy without a prescription. Some OTC medicines relieve aches, pains and itches. Some prevent or cure diseases, like tooth decay and athlete's foot. Others help manage recurring problems, like migraines.</code>
  <code>every how many weeks should you get a hair cut</code>	<code>“Somehow people have been taught you need to cut your hair every 4 to 6 weeks and I think that’s way too soon,” she tells InStyle. “If you have a great cut and don’t mind a little added length, the style can last up to 6 months and still look great.”. RELATED: We Tried It: Cindy Crawford’s Cleanse.</code>
  <code>how does a pedometer work</code>	<code>This is pretty much how a pedometer works. Photo: Pedometers can measure your steps because your body swings from side to side as you walk. Each swing counts as one step. Multiplying the number of swings by the average length of your steps tells you how far you've gone.</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

msmarco_pairs2

• Dataset: msmarco_pairs2 at 28ff31e
• Size: 6,876 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: 4 tokens</li><li>mean: 8.73 tokens</li><li>max: 25 tokens</li></ul>	<ul><li>min: 20 tokens</li><li>mean: 76.57 tokens</li><li>max: 268 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>what should house humidity percentage be?</code>	<code>Lower humidity can cause your heating system to work harder in the winter months. There must be a balance of humidity in your home to make it a comfortable place and preserve your interior as well as you heating and air conditioning systems. During summer months, homes should have between 30 and 45 percent humidity. During the winter, humidity levels should be kept between 45 and 55 percent humidity. How can I regulate my home humidity? The following are some ideas on how you can help to balance your home’s humidity: Install an inside weather station that measures humidity in your home.</code>
  <code>which electrolyte is missing if there is tetany</code>	<code>Hypocalcemia is not a term for tetany but is rather a cause of tetany. Causes. The usual cause of tetany is lack of calcium. An excess of phosphate (high phosphate-to-calcium ratio) can also trigger the spasms. Underfunction of the parathyroid gland can lead to tetany. Low levels of carbon dioxide cause tetany by altering the albumin binding of calcium such that the ionized (physiologically influencing) fraction of calcium is reduced; one common reason for low carbon dioxide levels is hyperventilation. Low levels of magnesium can lead to tetany.</code>
  <code>dimensions of the sea of galilee</code>	<code>Sea of Galilee From Wikipedia, the free encyclopedia. The Sea of Galilee is Israel's largest freshwater lake, approximately 53 kilometers (33 miles) in circumference, about 21 km (13 miles) long, and 13 km (8 miles) wide; it has a total area of 166 sq km, and a maximum depth of approximately 48 meters. At 213</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": 3,
      "last_layer_weight": 0.25,
      "prior_layers_weight": 2,
      "kl_div_weight": 0.75,
      "kl_temperature": 0.75
  }
  

nq_pairs

• Dataset: nq_pairs at f9e894e
• Size: 12,892 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: 10 tokens</li><li>mean: 11.88 tokens</li><li>max: 27 tokens</li></ul>	<ul><li>min: 7 tokens</li><li>mean: 130.01 tokens</li><li>max: 502 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>who did richmond play in the grand final</code>	<code>Richmond Football Club Richmond began 2017 with 5 straight wins, a feat it had not achieved since 1995. A series of close losses hampered the Tigers throughout the middle of the season, including a 5-point loss to the Western Bulldogs, 2-point loss to Fremantle, and a 3-point loss to the Giants. Richmond ended the season strongly with convincing victories over Fremantle and St Kilda in the final two rounds, elevating the club to 3rd on the ladder. Richmond's first final of the season – their qualifying final against the Cats at the MCG attracted a record qualifying final crowd of 95,028; the Tigers won by 51 points. In their first preliminary final since 2001, Richmond defeated Greater Western Sydney by 36 points in front of a crowd of 94,258 to progress to the Grand Final against Adelaide, their first Grand Final appearance since 1982. The attendance was 100,021, the largest crowd for a Grand Final since 1986. The Crows led at quarter time and led by as many as 13, but the Tigers took over the game as it progressed and kicked seven straight goals at one point. They eventually would win by 48 points – 16.12 (108) to Adelaide's 8.12 (60) – to end their 37-year flag drought.[23] Dustin Martin also became the first player to win a Premiership medal, the Brownlow Medal and the Norm Smith Medal in the same season, while Damien Hardwick was named AFL Coaches Association Coach of the Year. Richmond's jump from 13th to premiers also marked the biggest jump from one AFL season to the next.</code>
  <code>who was the first european to translate bhagwat gita into english</code>	<code>Charles Wilkins Sir Charles Wilkins, KH, FRS (1749 – 13 May 1836), was an English typographer and Orientalist, and founding member of The Asiatic Society. He is notable as the first translator of Bhagavad Gita into English, and as the creator, alongside Panchanan Karmakar,[1] of the first Bengali typeface.[2] In 1788, Wilkins was elected a member of the Royal Society.[3]</code>
  <code>who played bruce willis's girlfriend in pulp fiction</code>	<code>Maria de Medeiros Among Medeiros' most memorable film appearances are three early 1990s roles. Her considerable resemblance to Anaïs Nin landed her the primary role in Henry & June (1990), in which she played the author. In 1990, she played the role of Maria in Ken McMullen's film about the rise of the Paris Commune, 1871. In 1994, Medeiros appeared in Quentin Tarantino's Pulp Fiction playing Fabienne, the girlfriend of Butch Coolidge (Bruce Willis).</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": -1,
      "last_layer_weight": 2,
      "prior_layers_weight": 0.25,
      "kl_div_weight": 1.25,
      "kl_temperature": 0.9
  }
  

nq_pairs2

• Dataset: nq_pairs2 at f9e894e
• Size: 4,298 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: 10 tokens</li><li>mean: 11.75 tokens</li><li>max: 21 tokens</li></ul>	<ul><li>min: 14 tokens</li><li>mean: 127.94 tokens</li><li>max: 512 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>when does daylight saving time come into effect</code>	<code>Daylight saving time Start and end dates vary with location and year. Since 1996, European Summer Time has been observed from the last Sunday in March to the last Sunday in October; previously the rules were not uniform across the European Union.[39] Starting in 2007, most of the United States and Canada observe DST from the second Sunday in March to the first Sunday in November, almost two-thirds of the year.[43] The 2007 U.S. change was part of the Energy Policy Act of 2005; previously, from 1987 through 2006, the start and end dates were the first Sunday in April and the last Sunday in October, and Congress retains the right to go back to the previous dates now that an energy-consumption study has been done.[44] Proponents for permanently retaining November as the month for ending DST point to Halloween as a reason to delay the change—to provide extra daylight on October 31.</code>
  <code>when was the words under god added to the pledge of allegiance</code>	<code>Pledge of Allegiance (United States) The phrase "under God" was incorporated into the Pledge of Allegiance on June 14, 1954, by a Joint Resolution of Congress amending § 4 of the Flag Code enacted in 1942.[28]</code>
  <code>who provides the funds for a loan guaranteed by the veterans' administration</code>	<code>VA loan The basic intention of the VA home loan program is to supply home financing to eligible veterans and to help veterans purchase properties with no down payment. The loan may be issued by qualified lenders.</code>

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

  {
      "loss": "GISTEmbedLoss",
      "n_layers_per_step": 3,
      "last_layer_weight": 0.25,
      "prior_layers_weight": 2,
      "kl_div_weight": 0.75,
      "kl_temperature": 0.75
  }
  

trivia_pairs

• Dataset: trivia_pairs at a7c36e3
• Size: 17,190 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: 8 tokens</li><li>mean: 17.38 tokens</li><li>max: 62 tokens</li></ul>	<ul><li>min: 25 tokens</li><li>mean: 444.36 tokens</li><li>max: 512 tokens</li></ul>

• Samples:

  sentence1	sentence2
  <code>In which 2004 animated Pixar movie does Violet have powers of invisibility?</code>	<code>Violet - The Incredibles - Pixar movie - Character Profile - Writeups.org Jump to the game stats Content Violet is one of the main characters of the 2004 Pixar animation movie The Incredibles. The movie encountered both critical and commercial success. The Incredibles deals with super-heroism, family life and emotions, violence, and being different. Background Real Name: Violet Parr Other Aliases: Violet. It is very obvious from the film’s plot that a secret ID is incredibly important to a super’s life, so it is thus unlikely that Violet will continue to adventure as “Violet”. I suggest “Invisigirl” because it follows the film’s naming tradition of simple, descriptive names. Marital Status: Single Known Relatives: Robert Parr (Aka Mr. Incredible, father), Helen Parr (Aka Elastigirl, mother), Dashiell Robert Parr (Aka Dash, brother), Jack-Jack (Possibly Jacob) Parr (brother). Group Affiliation: The Incredibles Base Of Operations: Metroville Height: 5’ (Note:Violet’s official height is listed at 4’ 6”. but that is clearly not accurate, by comparing her height to her mother’s and to her boyfriend’s. I went with 5’, which seems accurate, but is still on the short side for a girl her age. Weight: 90lbs Age: 14 Eyes: Dark Blue Hair: Black (probably dyed) Powers and Abilities A pretty much normal teenaged girl otherwise, Violet/Invisigirl can become invisible (partially or fully) and can erect force fields with a high degree of imperviousness to harm (One of her fields protected her family from Syndrome’s crashing plane). Violet has difficulty using her Force Field when she is under stress. Video HD version of the official trailer. History Mr Incredible was his world’s most famed and lauded super-hero (supers, as they are called popularly), until a mishap while preventing a potential suicide led to a lawsuit for damages. This triggered not only an avalanche of personal-injury lawsuits against Mr Incredible personally, but a public backlash of opinion against supers in general, and most of them were forced to go underground to keep from being sued to death. Mr. Incredible and his new wife, Elastigirl , retired and became simple Mr. and Mrs. Parr, and started raising a family. Cut to 15 years later. Bob Parr is an insurance claims specialist with a midlife crisis and a desire to go back to the “old days”. He’s fed up with his pushy boss and his immoral profession. He and his best buddy Lucius Best (aka Frozone ), spend Wednesday nights cruising the city in a car, listening to the police scanner, and saving people on the sly. Helen is trying hard to be a mom to 3 kids, two of whom have superpowers of their own and fight constantly. She has worked too hard to build a normal life for her family to abide his nostalgia for heroism. Violet, their daughter, is having problems relating to people and is withdrawn and moody. Dash , their son, is chafing at the restrictions placed on him, and getting into trouble at school. When Mr. Incredible is offered the chance to play the role of hero again by a mysterious informant, he jumps at the opportunity. But when it turns out to be a trap set by an old nemesis he had a hand in corrupting, the whole family must reveal themselves to save Mr. Incredible and countless innocents. The Incredibles discover that their real power comes from their unity, rather than their superpowers. Description Violet is a tall, skinny girl with a retiring demeanor. She wears a duplicate of her mother’s Iincredisuit in red & black, with the Incredible’s logo on the chest. Before she came to terms with her “differentness”, she wore her hair so that it fell over her face, a figurative shield to hide behind, and wore mostly drab colors. She now wears it pulled back, and wears more bright colors. Personality Violet is a shy, worried teen who wants nothing more then to be normal. She grew up being taught to repress her abilities. Violet wants to be normal so badly that she has difficulty calling on her force field in anything but absolutely safe conditions… ie, at home, in the presence of her family. Her invisibility, a pur</code>
  <code>Where in the body is the Malleus bone?</code>	<code>Malleus Bone Definition, Function & Anatomy

The model card is truncated. Read the rest on Hugging Face.