hotchpotch's picture
Update README.md
f44b248 verified
metadata
language:
  - en
tags:
  - sentence-transformers
  - sentence-similarity
  - feature-extraction
  - dense
  - generated_from_trainer
  - dataset_size:4314846
  - loss:CachedMultipleNegativesBidirectionalRankingLoss
base_model: answerdotai/ModernBERT-base
widget:
  - source_sentence: what is grade 7 gcse equivalent to?
    sentences:
      - >-
        Unlike the Google Home Mini (First Gen), the Nest Mini (Second Gen) can
        be used to actually enjoy music in every room of the house. While the
        Google Home Mini (First Gen) is a decent way to get music in every room
        of your home for cheap, the sound quality that comes from the speaker
        reflects the price of the product.
      - >-
        In general, a grade 7-9 is roughly equivalent to A-A* under the old
        system, while a grade 4 and above is roughly equivalent to a C and
        above. Fewer students will receive a grade 9 than would have received an
        A* under the old grading system.
      - >-
        ['Pulling at a wet or dirty diaper.', 'Hiding to pee or poop.',
        "Interest in others' use of the potty, or copying their behavior.",
        'Having a dry diaper for a longer-than-usual time.', 'Awakening dry from
        a nap.', "Telling you that they're about to go, are going or have just
        gone in their diaper."]
  - source_sentence: >-
      Desire For Sex Drops As You Age, But You Can Still Have A Satisfactory Sex
      Life
    sentences:
      - >-
        ADVERTISEMENT

        Those who have been in long-term relationships know that sex can start
        to fall by the wayside the longer you're together.

        Whether you have children, a busy career, an active social life, a job
        that takes you away from home often, or a chronic illness, there are
        plenty of reasons why couples have less sex compared to when they first
        started dating.

        And it's not just stuff like that that's keeping you away from fun
        between the sheets; according to research from the Kinsey Institute, age
        plays a factor in your sex drive, for both men and women.

        Unsurprisingly, younger people are having the most sex compared to other
        age groups.

        Those aged 18 to 29 years old are having sex an average of 112 times a
        year (about every three days), and, as Indy100 notes, most people lose
        their virginity when they're teenagers, with men having sex for the
        first time around 16.8 years, and women losing theirs at 17.2 years.

        By comparison, 30 to 39-year-olds have sex on average 86 times a year,
        which is around 1.6 times per week.

        The study notes that this drop-off coincides with the age people choose
        to start having children, which, as parents know, can really kill the
        mood, especially if there's a baby crying at the exact same time you
        feel like getting it on. (Which is most likely in the morning.)

        And it only lessens the older you get. Those who are in their 40s have
        sex an average of 69 times a year, due to factors such as family
        obligations, day-to-day stresses, and possible illnesses.

        "The basic storyline that has emerged from these studies is that, as we
        get older, our odds of developing chronic health conditions increases
        and this, in turn, negatively impacts the frequency and quality of
        sexual activity," notes Dr. Justin Lehmiller of the Kinsey Institute.

        Unfortunately, the study didn't look into the sex lives of those 50 and
        older, but there is other research out there. According to a study
        published in the Archives of Sexual Behavior, couples who have been
        married for more than 25 years have a 40 per cent chance of having sex
        two or three times a week, but that statistic drops to 35 per cent for
        couples who have been married for 50 or more years.

        Surprisingly, couples who have been together for 65 years are 42 per
        cent more likely to have sex a couple times a week.

        As we get older, our odds of developing chronic health conditions
        increases and this, in turn, negatively impacts the frequency and
        quality of sexual activity.

        According to a study published in the Journal of Sex Research, those who
        "feel their age" tended to have less sex, while those who remained in
        better health had more active and satisfying sex lives.

        "The younger people feel, the more likely they are to maintain high
        sexual satisfaction as they get older (or at least they'll experience a
        much less noticeable change)," wrote Lehmiller.

        It's worth noting that these study results come from a small sample of
        the population, and it shouldn't be the standard for how much sex we
        should be having.

        However, there is plenty of research that backs up the claim that sex is
        great for one's health, so the more you get busy, the better!

        Also on HuffPost:
      - >-
        HONOLULU — A former Hawaii state worker who sent a false missile alert
        last month said Friday that he's devastated for causing panic but was
        "100 per cent sure" at the time that the attack was real.

        The man in his 50s spoke to reporters on the condition that he not be
        identified because he fears for his safety after receiving threats.

        He says the on-duty call he received on Jan. 13 didn't sound like a
        drill. However, state officials say other workers clearly heard the word
        "exercise" repeated several times.

        He said it felt like he had been hit with a "body blow" when he realized
        it was just a drill and he has had difficulty eating and sleeping since.

        The Hawaii Emergency Management Agency fired him.

        The man's superiors said they knew for years that he had problems
        performing his job. The worker had mistakenly believed drills for
        tsunami and fire warnings were actual events, and colleagues were not
        comfortable working with him, the state said.

        His supervisors counselled him but kept him for a decade in a position
        that had to be renewed each year.

        The ex-worker disputed that, saying he wasn't aware of any performance
        problems.

        While working at the state warning site in a former bunker in Honolulu's
        Diamond Head crater on Jan. 13, the man said, he took a call that
        sounded like a real warning from U.S Pacific Command. He said he didn't
        hear that it was a drill.

        But the problems at the agency went beyond the one employee.

        Federal and state reports say the agency had a vague checklist for
        missile alerts, allowing workers to interpret the steps they should
        follow differently. Managers didn't require a second person to sign off
        on alerts before they were sent, and the agency lacked any preparation
        on how to correct a false warning.

        Those details emerged Tuesday in reports on investigations about how the
        agency mistakenly blasted cellphones and broadcast stations with the
        missile warning.

        It took nearly 40 minutes for the agency to figure out a way to retract
        the false alert on the same platforms it was sent to.

        "The protocols were not in place. It was a sense of urgency to put it in
        place as soon as possible. But those protocols were not developed to the
        point they should have," retired Brig. Gen. Bruce Oliveira, who wrote
        the report on Hawaii's internal investigation, said at a news
        conference.

        Hawaii Emergency Management Agency Administrator Vern Miyagi resigned as
        the reports were released. Officials revealed that the employee who sent
        the alert was fired Jan. 26. The state did not name him.

        The agency's executive officer, Toby Clairmont, said Wednesday that he
        stepped down because it was clear action would be taken against agency
        leaders after the alert.
      - >-
        Pompeii’s Final Hours: New Evidence (C5)

        Rating:

        The Big Crash Diet Experiment (BBC1)

        Rating:

        With his rosy cheeks and nose, and a crown of laurel leaves drooping
        over one eye, former political journalist John Sergeant looked like
        jolly little Bacchus, the Roman god of wine, as he tucked into an
        ancient feast on Pompeii’s Final Hours: New Evidence (C5).

        A game soul, whether strutting the pasa doble on Strictly or bartering
        in a Naples marketplace, John munched fried sea urchins and braised
        moray eel  with plenty of red vino to slosh the taste away.

        He did blanch at the thought of bulls’ testicles stuffed with pepper and
        herbs.

        John Sergeant on an hour-long archaeological romp in Pompeii’s Final
        Hours: New Evidence

        Apparently this delicacy was a great favourite in Pompeii  but then,
        the decadent Romans drenched every meal in lashings of garum, a sauce
        made from rotting fish. Anything would taste better than that.

        Noble as Brutus, John held his nose and chewed a mouthful of cobbler. ‘I
        wouldn’t have it every night,’ he muttered.

        It’s an astonishing thought that Julius Caesar conquered most of the
        known world, when he must have been suffering from chronic indigestion.

        Imagine what the Romans might have done if they’d invented the pizza a
        couple of thousand years earlier.

        This hour-long archaeological romp was the first of three surveys of
        life in the shadow of Vesuvius, set to continue tonight and tomorrow.

        The ‘new evidence’ in the title came from computer X-ray scans of some
        of Pompeii’s famous casts.

        These detailed figurines were created by the 19th-century archaeologist
        Giuseppe Fiorelli, who injected liquid plaster into the cavities where
        Roman bodies had been buried by ash in the volcanic eruption in AD79.

        Fiorelli’s casts are the most moving and tragic death masks ever made.
        Every plaster corpse is writhing in agony, suffocated by poisonous
        gases.

        For 150 years, the victims’ skeletal remains have been locked in their
        cases. It is only now that the technology exists to examine the bones
        without destroying the casts.

        What the first CT scans revealed swept old theories away. One figure
        long believed to be a man appeared, in fact, to be female.

        Another, thought for decades to be a male gladiator in his prime, turned
        out to be a teenage boy.

        Presenters Bettany Hughes and Raksha Dave didn’t make enough of these
        dramatic finds. The CT results were held back to the end of the hour, so
        that the discoveries were inevitably rushed.

        Dr Javid Abdelmoneim in The Big Crash Diet Experiment challenges
        conventional wisdom on food and exercise

        Don’t blame John Sergeant, though. While the others were in the lab, he
        was still polishing off his meal of eels and urchins. Say what you like,
        this man believes in doing his research.

        After that, he’d probably welcome a few days of starvation. The powdered
        soups and shakes fed to four slimmers by Dr Javid Abdelmoneim in The Big
        Crash Diet Experiment (BBC1) looked worse than any classical culinary
        torture, though.

        To challenge conventional wisdom that brief bursts of intensive dieting
        rarely bring long-term results, Dr Javid had his guinea pigs living on
        800 calories a day for nine weeks.

        All lost plenty of weight. But it was the switch to healthy-eating
        afterwards that seemed to bring the best results.

        The show had plenty of useful advice for dieters. Don’t pretend fast
        food is ‘addictive’  greasy take-aways are just a bad habit. Only eat
        in the dining room, never on the sofa . . . or in bed.

        Remember, burger bars are in the cynical business of selling you empty
        calories.

        Follow those rules, and you might not need the powdered shakes. Or the
        foul fish sauce.
  - source_sentence: Berlin startup offers a year with no money worries
    sentences:
      - >-
        Get daily updates directly to your inbox + Subscribe Thank you for
        subscribing! Could not subscribe, try again later Invalid Email

        Nuneaton's hospital has been given the all-clear after a previously
        closed ward has now been re-opened.

        Bosses at the George Eliot Hospital were forced to close the Adam Bede
        ward due to an outbreak of Norovirus.

        It remained closed over the weekend but on Monday they said that ward
        had now been decontaminated and re-opened.

        Martina Morris, deputy director of nursing at George Eliot Hospital NHS
        Trust, said: “The patients on Adam Bede ward have been clear of symptoms
        for the last 48 hours, and following a full decontamination, we have
        re-opened the ward.

        “Any patients in the hospital who continue to present with symptoms of
        norovirus have been isolated in side rooms.”

        “But they are keen to prevent any further outbreaks and are appealing to
        anyone from suffering from the sickness and diarrhoea to steer clear.

        “We ask that the public continue to avoid the hospital, if they have
        symptoms of diarrhoea and vomiting and do not visit until they have been
        symptom free for at least 48 hours,” the deputy director of nursing
        said.

        “Good hand hygiene is key to limiting the spread of these infections and
        it is important to wash your hands thoroughly with soap and warm water
        as using just an anti-bacterial hand gel is not sufficient.”
      - >-
        Comedy cabaret team All That Malarkey are promising to end 2017 with a
        festive bang with their new show Camp as Christmas.

        They will be playing The Groundlings Theatre in Portsmouth on December
        20 at 7.30pm (www.groundlings.co.uk) and Chichester’s St John’s Chapel
        on December 21, also at 7.30pm (07722 824696).

        Spokesman David Harrington said: “We spent a sizzling summer strutting
        our stuff at the Edinburgh Fringe Festival, where we performed to an
        international audience and gained excellent reviews.”

        Now they are back on the road for Christmas: “We’re excited to have
        dates including our London debut at the magnificent King’s Head Theatre,
        as well as other performances in Wales and the South, though we always
        finish at Chichester as that is where our journey began.

        “The four classically-trained singers of ATM are geared up and ready to
        sing their hearts out, fling themselves around the stage and present
        popular Christmas songs from pop to classics and carols, all musically
        arranged in unexpected ways that will surprise and entertain,
        accompanied and compered by yours truly at the keyboard. Known for our
        unique four-part harmony arrangements of family favourites, laced with
        fun, sparkle and tongue-in-cheek frivolity, our new programme will
        include wonderful new renditions of Do you ACTUALLY wish it could be
        Christmas everyday, Christmas No.1 Medley and We Need a Little
        Christmas.

        “Always drawing an amazing and welcoming crowd, our performance this
        year will be at St John’s Chapel, Chichester, hometown of the unmissable
        ginger-haired ATM soprano, Amy Fuller, and the city where ATM started
        four Christmases ago.

        “Promising to be an energetic and impossibly-festive evening, we’ll also
        be holding a collection for St Wilfrid’s Hospice at the end,
        particularly close to our hearts this year. Also in the diary for this
        tour is an appearance at my hometown of Portsmouth (Wednesday, December
        20 at The Groundlings Theatre). Having gone to Padnell school and
        Oaklands Catholic school and sixth form, it will be a treat to bring our
        outrageous act to old friends and family, and show them what I do for a
        living…flick my hair around and make funny faces at the piano like a
        maniac. Amy Fuller had made herself a complete stranger to me by growing
        up in Chichester and going to Bishop Luffa and Parklands Primary, but we
        fortunately crossed paths when studying together.”
      - >-
        Michael Bohmeyer, the founder of Mein Grundeinkommen (My Basic Income).
        Photo: DPA

        Miko from Berlin may only be five, but he already has €1,000 ($1,063)
        per month to live on -- not from hard graft, but as part of an
        experiment into universal basic income.

        He is one of 85 people, including around 10 children, chosen by startup
        Mein Grundeinkommen (My Basic Income) to receive the payments for a year
        since 2014.

        Founder Michael Bohmeyer has set out to prove to a sceptical public in
        Germany and further afield that the universal basic income (UBI) idea is
        workable.

        "Thanks to my first startup, I got a regular income, my life became more
        creative and healthy. So I wanted to launch a social experiment,"
        31-year-old Bohmeyer told AFP.

        And he wasn't alone in wanting to test the idea, as some 55,000 donors
        have stumped up the cash for the payments in a "crowdfunding" model --
        with the final recipients picked out in a "wheel of fortune" event
        livestreamed online.

        Mother Birgit Kaulfuss said little Miko "can't really understand, but
        for the whole family it was exhilarating" when he was chosen -- offering
        a chance to live "in a more relaxed way" and take a first-ever family
        holiday.

        Trying things out

        "Everyone sleeps more soundly and no one become a layabout," Bohmeyer
        said of his beneficiaries.

        Recipients' experiences range from a welcome spell without financial
        worries to major turning points in their lives.

        "Without day-to-day pressures, you can be more creative and try things
        out," Valerie Rupp told public broadcaster ARD in a recent interview.

        She was able both to take care of her baby and start a career as a
        decorator -- even as her husband, newly arrived from Mali, was taking
        German

        lessons.

        Winners have left jobs that were doing little more for them than put
        bread on the table to become teachers, taken time out to address chronic
        illness, broken alcohol addiction, taken care of loved ones, or paid for
        children's studies.

        "It's at once a gift and a prompt" to make a change, explained Astrid
        Lobeyer, who used the money to give eulogies at funerals and studied the

        therapeutic Alexander technique, a method for relieving stress in the
        muscles.

        Bohmeyer's experiment has fascinated social media and boosted discussion
        about a universal income in Germany.

        At the same time, Finland is testing the idea with 2,000 homeless
        recipients and the idea is a flagship policy for French Socialist
        presidential

        candidate Benoit Hamon.

        Reward for laziness?

        In 2009, the German parliament flatly rejected a petition from some
        50,000 Germans demanding a universal income.

        Nevertheless, some 40 percent of the public still think it's a good
        idea, according to a survey last June by pollsters Emnid.

        Supporters have formed a campaign group called "Buendnis Grundeinkommen"
        (Basic income federation) with their sights on September's legislative
        elections, but so far no major party has taken up the cause.

        There are pockets of support among left-wingers, the right, Catholic
        organisations and even industry leaders, whose reasoning ranges from
        fighting poverty to simplifying bureaucracy or smoothing the transition
        into the

        digital era.

        Resistance to the idea is more focused, centering on how UBI would
        change people's relationship to work.

        Right-wingers dismiss it as a "reward for laziness", while the Social
        Democratic Party (SPD) worried in 2006 about unemployed recipients being

        "labelled useless" rather than getting help to find jobs.

        Meanwhile, major unions like IG Metall and Verdi denounce the idea as a
        "liberal Trojan horse" that would "boost inequality" by paying
        millionaires and poor people alike.

        Thankless jobs

        Mein Grundeinkommen is "poorly thought out" as a response to broader
        social questions, University of Freiburg economist Alexander Spermann
        told AFP.

        The startup's 20 employees eat up "60 percent of the budget", founder
        Michael Bohmeyer admits -- while the idea of basing the funding on
        curiosity or activism by thousands of donors is hardly applicable on a
        large scale.

        For Spermann, the Berliners' experiment has only succeeded in answering
        the question "what would I do with a blank cheque if I got one for
        Christmas?"

        People's choices in terms of qualifications or work if they were
        guaranteed the payments for life are the real mystery, the economist
        argues.

        "Who will take on the exhausting and sometimes less attractive tasks,
        like emptying bins or taking care of the elderly?" asked Werner
        Eichhorst of the Bonn Centre for the Future of Work (IZA) in 2013.

        UBI supporters argue such jobs would either be taken over by robots or
        find a new place of honour in society if the policy were enacted.

        "No machine will take over working for us and pay our taxes at the same
        time," Eichhorst and opponents shoot back.
  - source_sentence: population of artesia
    sentences:
      - >-
        Meanwhile, bring 4 cups of water to a boil and add the barley. Simmer
        uncovered for 30 minutes, drain, and set aside. When the soup is ready,
        add the barley and cook the soup for another 15 or 20 minutes, until the
        barley is tender.
      - >-
        The 2016 Artesia, New Mexico, population is 12,036. There are 1,211
        people per square mile (population density).
      - >-
        There are 30 calories in one cup of chopped green peppers and
        approximately 6 calories in 1 ounce or 28g of green peppers.
  - source_sentence: what is the best paying engineering job
    sentences:
      - >-
        The 20 highest-paying jobs for engineering majors. Engineering jobs pay
        well. To find out just how lucrative they really are, we turned to
        PayScale, the creator of the world's largest compensation database. To
        find the 20 highest-paying jobs for engineering majors, PayScale first
        identified the most common jobs for those with a bachelor's degree (and
        nothing more) who work full-time in the US. Chief architects and vice
        president's of business development topped the list, both earning an
        impressive $151,000 a year.
      - "Depending on the thickness and size of the chop, it can take anywhere from eight to 30 minutes. Hereâ\x80\x99s a helpful cooking chart and some tips to achieve delicious pork chops every time. Pork chops are a crowd pleaser, especially once you master your grilling technique. For safe consumption, itâ\x80\x99s recommended to cook pork until it reaches an internal temperature of 145°F or 65°C. Depending on the cut and thickness of your chop, the time it may take to reach this can vary. To make sure your chops are the right temperature, use a digital meat thermometer."
      - >-
        Aviation is a combat arms branch which encompasses 80 percent of the
        commissioned officer operational flying positions within the Army (less
        those in Aviation Material Management and Medical Service Corps).
datasets:
  - sentence-transformers/msmarco-co-condenser-margin-mse-sym-mnrl-mean-v1
  - sentence-transformers/natural-questions
  - sentence-transformers/gooaq
  - sentence-transformers/ccnews
  - sentence-transformers/hotpotqa
pipeline_tag: sentence-similarity
library_name: sentence-transformers
metrics:
  - cosine_accuracy@10
  - cosine_precision@10
  - cosine_recall@10
  - cosine_ndcg@10
  - cosine_mrr@10
  - cosine_map@10
model-index:
  - name: SentenceTransformer based on answerdotai/ModernBERT-base
    results:
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoClimateFEVER
          type: NanoClimateFEVER
        metrics:
          - type: cosine_accuracy@10
            value: 0.68
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.09
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.374
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.3204103646278264
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.42072222222222216
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.2384825396825397
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoDBPedia
          type: NanoDBPedia
        metrics:
          - type: cosine_accuracy@10
            value: 0.94
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.39000000000000007
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.2684345324233032
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.5013173913967965
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.7496666666666667
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.3713051587301587
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoFEVER
          type: NanoFEVER
        metrics:
          - type: cosine_accuracy@10
            value: 0.98
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.10199999999999998
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.9333333333333332
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.7970708195176515
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.7731666666666667
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.7398333333333332
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoFiQA2018
          type: NanoFiQA2018
        metrics:
          - type: cosine_accuracy@10
            value: 0.74
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.122
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.5628492063492063
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.45952453703882723
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.5142222222222222
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.3760648589065255
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoHotpotQA
          type: NanoHotpotQA
        metrics:
          - type: cosine_accuracy@10
            value: 0.94
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.12999999999999998
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.65
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6496205965616751
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.8305555555555556
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.5639444444444445
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoMSMARCO
          type: NanoMSMARCO
        metrics:
          - type: cosine_accuracy@10
            value: 0.84
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.08399999999999999
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.84
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.5914940146382726
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.5118333333333333
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.5118333333333334
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoNFCorpus
          type: NanoNFCorpus
        metrics:
          - type: cosine_accuracy@10
            value: 0.7
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.256
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.13512669313971043
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.29812924809751384
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.4497777777777777
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.20484007936507936
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoNQ
          type: NanoNQ
        metrics:
          - type: cosine_accuracy@10
            value: 0.78
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.08399999999999999
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.76
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6278509641999098
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.5953333333333333
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.5760000000000001
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoQuoraRetrieval
          type: NanoQuoraRetrieval
        metrics:
          - type: cosine_accuracy@10
            value: 1
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.132
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.986
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.9386568522919021
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.9366666666666665
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.9120888888888888
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoSCIDOCS
          type: NanoSCIDOCS
        metrics:
          - type: cosine_accuracy@10
            value: 0.82
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.176
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.35966666666666663
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.3412893142888829
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.5091904761904761
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.23048174603174598
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoArguAna
          type: NanoArguAna
        metrics:
          - type: cosine_accuracy@10
            value: 0.9
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.09
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.9
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.589790277339453
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.49080158730158724
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.4908015873015873
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoSciFact
          type: NanoSciFact
        metrics:
          - type: cosine_accuracy@10
            value: 0.8
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.092
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.8
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.6514145845317466
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6098333333333332
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.5992222222222222
            name: Cosine Map@10
      - task:
          type: information-retrieval
          name: Information Retrieval
        dataset:
          name: NanoTouche2020
          type: NanoTouche2020
        metrics:
          - type: cosine_accuracy@10
            value: 0.9387755102040817
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.4102040816326531
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.2819732491937568
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.4762218106016415
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.720262390670554
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.3260029000262236
            name: Cosine Map@10
      - task:
          type: nano-beir
          name: Nano BEIR
        dataset:
          name: NanoBEIR mean
          type: NanoBEIR_mean
        metrics:
          - type: cosine_accuracy@10
            value: 0.8506750392464679
            name: Cosine Accuracy@10
          - type: cosine_precision@10
            value: 0.16601569858712717
            name: Cosine Precision@10
          - type: cosine_recall@10
            value: 0.603952590854306
            name: Cosine Recall@10
          - type: cosine_ndcg@10
            value: 0.5571377519332383
            name: Cosine Ndcg@10
          - type: cosine_mrr@10
            value: 0.6240024793800304
            name: Cosine Mrr@10
          - type: cosine_map@10
            value: 0.4723770070973909
            name: Cosine Map@10
license: apache-2.0

SentenceTransformer based on answerdotai/ModernBERT-base

This is a sentence-transformers model finetuned from answerdotai/ModernBERT-base on the msmarco, natural_questions, gooaq, ccnews and hotpotqa 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 Sources

Full Model Architecture

SentenceTransformer(
  (0): Transformer({'max_seq_length': 512, 'do_lower_case': False, 'architecture': 'ModernBertModel'})
  (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("hotchpotch/ModernBERT-embedding-CMNBRL")
# Run inference
queries = [
    "what is the best paying engineering job",
]
documents = [
    "The 20 highest-paying jobs for engineering majors. Engineering jobs pay well. To find out just how lucrative they really are, we turned to PayScale, the creator of the world's largest compensation database. To find the 20 highest-paying jobs for engineering majors, PayScale first identified the most common jobs for those with a bachelor's degree (and nothing more) who work full-time in the US. Chief architects and vice president's of business development topped the list, both earning an impressive $151,000 a year.",
    'Aviation is a combat arms branch which encompasses 80 percent of the commissioned officer operational flying positions within the Army (less those in Aviation Material Management and Medical Service Corps).',
    'Depending on the thickness and size of the chop, it can take anywhere from eight to 30 minutes. Hereâ\x80\x99s a helpful cooking chart and some tips to achieve delicious pork chops every time. Pork chops are a crowd pleaser, especially once you master your grilling technique. For safe consumption, itâ\x80\x99s recommended to cook pork until it reaches an internal temperature of 145°F or 65°C. Depending on the cut and thickness of your chop, the time it may take to reach this can vary. To make sure your chops are the right temperature, use a digital meat thermometer.',
]
query_embeddings = model.encode_query(queries)
document_embeddings = model.encode_document(documents)
print(query_embeddings.shape, document_embeddings.shape)
# [1, 768] [3, 768]

# Get the similarity scores for the embeddings
similarities = model.similarity(query_embeddings, document_embeddings)
print(similarities)
# tensor([[0.9709, 0.7909, 0.6977]])

Evaluation

Metrics

Information Retrieval

  • Datasets: NanoClimateFEVER, NanoDBPedia, NanoFEVER, NanoFiQA2018, NanoHotpotQA, NanoMSMARCO, NanoNFCorpus, NanoNQ, NanoQuoraRetrieval, NanoSCIDOCS, NanoArguAna, NanoSciFact and NanoTouche2020
  • Evaluated with InformationRetrievalEvaluator
Metric NanoClimateFEVER NanoDBPedia NanoFEVER NanoFiQA2018 NanoHotpotQA NanoMSMARCO NanoNFCorpus NanoNQ NanoQuoraRetrieval NanoSCIDOCS NanoArguAna NanoSciFact NanoTouche2020
cosine_accuracy@10 0.68 0.94 0.98 0.74 0.94 0.84 0.7 0.78 1.0 0.82 0.9 0.8 0.9388
cosine_precision@10 0.09 0.39 0.102 0.122 0.13 0.084 0.256 0.084 0.132 0.176 0.09 0.092 0.4102
cosine_recall@10 0.374 0.2684 0.9333 0.5628 0.65 0.84 0.1351 0.76 0.986 0.3597 0.9 0.8 0.282
cosine_ndcg@10 0.3204 0.5013 0.7971 0.4595 0.6496 0.5915 0.2981 0.6279 0.9387 0.3413 0.5898 0.6514 0.4762
cosine_mrr@10 0.4207 0.7497 0.7732 0.5142 0.8306 0.5118 0.4498 0.5953 0.9367 0.5092 0.4908 0.6098 0.7203
cosine_map@10 0.2385 0.3713 0.7398 0.3761 0.5639 0.5118 0.2048 0.576 0.9121 0.2305 0.4908 0.5992 0.326

Nano BEIR

  • Dataset: NanoBEIR_mean
  • Evaluated with NanoBEIREvaluator with these parameters:
    {
        "dataset_names": [
            "climatefever",
            "dbpedia",
            "fever",
            "fiqa2018",
            "hotpotqa",
            "msmarco",
            "nfcorpus",
            "nq",
            "quoraretrieval",
            "scidocs",
            "arguana",
            "scifact",
            "touche2020"
        ],
        "dataset_id": "sentence-transformers/NanoBEIR-en"
    }
    
Metric Value
cosine_accuracy@10 0.8507
cosine_precision@10 0.166
cosine_recall@10 0.604
cosine_ndcg@10 0.5571
cosine_mrr@10 0.624
cosine_map@10 0.4724

Training Details

Training Datasets

msmarco

msmarco

  • Dataset: msmarco at 84ed2d3
  • Size: 502,939 training samples
  • Columns: query and positive
  • Approximate statistics based on the first 1000 samples:
    query positive
    type string string
    details
    • min: 4 tokens
    • mean: 9.26 tokens
    • max: 25 tokens
    • min: 19 tokens
    • mean: 80.68 tokens
    • max: 230 tokens
  • Samples:
    query positive
    is cabinet refacing worth the cost? Fans of refacing say this mini-makeover can give a kitchen a whole new look at a much lower cost than installing all-new cabinets. Cabinet refacing can save up to 50 percent compared to the cost of replacing, says Cheryl Catalano, owner of Kitchen Solvers, a cabinet refacing franchise in Napierville, Illinois. From.
    is the fovea ethmoidalis a bone Ethmoid bone/fovea ethmoidalis. The medial portion of the ethmoid bone is a cruciate membranous bone composed of the crista galli, cribriform plate, and perpendicular ethmoidal plate. The crista is a thick piece of bone, shaped like a “cock's comb,” that projects intracranially and attaches to the falx cerebri.
    average pitches per inning The likelihood of a pitcher completing nine innings if he throws an average of 14 pitches or less per inning is reinforced by the totals of the 89 games in which pitchers did actually complete nine innings of work.
  • Loss: CachedMultipleNegativesBidirectionalRankingLoss with these parameters:
    {
        "temperature": 0.01,
        "similarity_fct": "cos_sim",
        "mini_batch_size": 128,
        "gather_across_devices": false
    }
    
natural_questions

natural_questions

  • Dataset: natural_questions at f9e894e
  • Size: 100,231 training samples
  • Columns: query and positive
  • Approximate statistics based on the first 1000 samples:
    query positive
    type string string
    details
    • min: 10 tokens
    • mean: 12.46 tokens
    • max: 22 tokens
    • min: 12 tokens
    • mean: 137.8 tokens
    • max: 512 tokens
  • Samples:
    query positive
    difference between russian blue and british blue cat Russian Blue The coat is known as a "double coat", with the undercoat being soft, downy and equal in length to the guard hairs, which are an even blue with silver tips. However, the tail may have a few very dull, almost unnoticeable stripes. The coat is described as thick, plush and soft to the touch. The feeling is softer than the softest silk. The silver tips give the coat a shimmering appearance. Its eyes are almost always a dark and vivid green. Any white patches of fur or yellow eyes in adulthood are seen as flaws in show cats.[3] Russian Blues should not be confused with British Blues (which are not a distinct breed, but rather a British Shorthair with a blue coat as the British Shorthair breed itself comes in a wide variety of colors and patterns), nor the Chartreux or Korat which are two other naturally occurring breeds of blue cats, although they have similar traits.
    who played the little girl on mrs doubtfire Mara Wilson Mara Elizabeth Wilson[2] (born July 24, 1987) is an American writer and former child actress. She is known for playing Natalie Hillard in Mrs. Doubtfire (1993), Susan Walker in Miracle on 34th Street (1994), Matilda Wormwood in Matilda (1996) and Lily Stone in Thomas and the Magic Railroad (2000). Since retiring from film acting, Wilson has focused on writing.
    what year did the movie the sound of music come out The Sound of Music (film) The film was released on March 2, 1965 in the United States, initially as a limited roadshow theatrical release. Although critical response to the film was widely mixed, the film was a major commercial success, becoming the number one box office movie after four weeks, and the highest-grossing film of 1965. By November 1966, The Sound of Music had become the highest-grossing film of all-time—surpassing Gone with the Wind—and held that distinction for five years. The film was just as popular throughout the world, breaking previous box-office records in twenty-nine countries. Following an initial theatrical release that lasted four and a half years, and two successful re-releases, the film sold 283 million admissions worldwide and earned a total worldwide gross of $286,000,000.
  • Loss: CachedMultipleNegativesBidirectionalRankingLoss with these parameters:
    {
        "temperature": 0.01,
        "similarity_fct": "cos_sim",
        "mini_batch_size": 128,
        "gather_across_devices": false
    }
    
gooaq

gooaq

  • Dataset: gooaq at b089f72
  • Size: 3,012,496 training samples
  • Columns: query and positive
  • Approximate statistics based on the first 1000 samples:
    query positive
    type string string
    details
    • min: 8 tokens
    • mean: 12.05 tokens
    • max: 21 tokens
    • min: 13 tokens
    • mean: 59.08 tokens
    • max: 116 tokens
  • Samples:
    query positive
    how do i program my directv remote with my tv? ['Press MENU on your remote.', 'Select Settings & Help > Settings > Remote Control > Program Remote.', 'Choose the device (TV, audio, DVD) you wish to program. ... ', 'Follow the on-screen prompts to complete programming.']
    are rodrigues fruit bats nocturnal? Before its numbers were threatened by habitat destruction, storms, and hunting, some of those groups could number 500 or more members. Sunrise, sunset. Rodrigues fruit bats are most active at dawn, at dusk, and at night.
    why does your heart rate increase during exercise bbc bitesize? During exercise there is an increase in physical activity and muscle cells respire more than they do when the body is at rest. The heart rate increases during exercise. The rate and depth of breathing increases - this makes sure that more oxygen is absorbed into the blood, and more carbon dioxide is removed from it.
  • Loss: CachedMultipleNegativesBidirectionalRankingLoss with these parameters:
    {
        "temperature": 0.01,
        "similarity_fct": "cos_sim",
        "mini_batch_size": 128,
        "gather_across_devices": false
    }
    
ccnews

ccnews

  • Dataset: ccnews at 6118cc0
  • Size: 614,664 training samples
  • Columns: query and positive
  • Approximate statistics based on the first 1000 samples:
    query positive
    type string string
    details
    • min: 7 tokens
    • mean: 16.71 tokens
    • max: 56 tokens
    • min: 18 tokens
    • mean: 349.3 tokens
    • max: 512 tokens
  • Samples:
    query positive
    Rupee rises for 2nd consecutive day, gains 8 paise against US dollar today The rupee rose 8 paise to close at 64.37 apiece US dollar at the interbank foreign exchange market today.
    The Indian rupee appreciated for the second consecutive day and gained over 8 paise against the US dollar on Monday. The domestic currency opened unchanged today, very quickly edged higher and extended the gains to hit a day’s high of 64.34. The rupee rose 8 paise to close at 64.37 apiece US dollar at the interbank foreign exchange market today. The Reserve Bank of India fixed the reference rate of the rupee at 64.3616 against the US dollar on Monday. The Indian rupee moved up 23 paise against the US dollar in just 2 days as Narendra Modi led BJP is most likely to conquer Gujarat for the fifth consecutive time in the state elections. Way back in March 2017, the rupee appreciated as much as 79 paise in a single day to close at a 16-month high against the US dollar after Bharatiya Janata Party’s landslide victory in Uttar Pradesh state elections.
    Finance Minister Arun Jaitley is all ...
    Microsoft pushes for ‘Digital Geneva Convention’ for cybercrimes Technology companies, he added, need to preserve trust and stability online by pledging neutrality in cyber conflict. ( Image for representation, Source: Reuters) Technology companies, he added, need to preserve trust and stability online by pledging neutrality in cyber conflict. ( Image for representation, Source: Reuters)
    Microsoft President Brad Smith on Tuesday pressed the world’s governments to form an international body to protect civilians from state-sponsored hacking, saying recent high-profile attacks showed a need for global norms to police government activity in cyberspace.
    Countries need to develop and abide by global rules for cyber attacks similar to those established for armed conflict at the 1949 Geneva Convention that followed World War Two, Smith said. Technology companies, he added, need to preserve trust and stability online by pledging neutrality in cyber conflict.
    Watch all our videos from Express Technology
    “We need a Digital Geneva Convention that will commit go...
    Prince Gets Purple Pantone Color ‘Love Symbol #2’ By Abby Hassler
    Prince, also known as “The Purple One” is finally getting his very own Pantone color. Pantone and Prince’s Estate announced today (August 14) that the late singer has his own purple hue, “Love Symbol #2,” which is named after the iconic symbol the singer used as an emblem for his name.
    Related: Wesley Snipes Beat Out Prince for His Role in Michael Jackson’s ‘Bad’
    “The color purple was synonymous with who Prince was and will always be. This is an incredible way for his legacy to live on forever,” Troy Carter, entertainment adviser to Prince’s Estate, said.
    “We are honored to have worked on the development of Love Symbol #2, a distinctive new purple shade created in memory of Prince, ‘the purple one,'” added Laurie Pressman, vice president of the Pantone Color Institute. “A musical icon known for his artistic brilliance, Love Symbol #2 is emblematic of Prince’s distinctive style. Long associated with the purple family, Love Symbol #2 enables Prince’s unique purple shade t...
  • Loss: CachedMultipleNegativesBidirectionalRankingLoss with these parameters:
    {
        "temperature": 0.01,
        "similarity_fct": "cos_sim",
        "mini_batch_size": 128,
        "gather_across_devices": false
    }
    
hotpotqa

hotpotqa

  • Dataset: hotpotqa at f07d3cd
  • Size: 84,516 training samples
  • Columns: query and positive
  • Approximate statistics based on the first 1000 samples:
    query positive
    type string string
    details
    • min: 8 tokens
    • mean: 25.82 tokens
    • max: 140 tokens
    • min: 18 tokens
    • mean: 103.34 tokens
    • max: 350 tokens
  • Samples:
    query positive
    Which magazine covers a wider range of topics, Decibel or Paper? Decibel (magazine) Decibel is a monthly heavy metal magazine published by the Philadelphia-based Red Flag Media since October 2004. Its sections include Upfront, Features, Reviews, Guest Columns and the Decibel Hall of Fame. The magazine's tag-line is currently "Extremely Extreme" (previously "The New Noise"); the editor-in-chief is Albert Mudrian.
    what bbc drama features such actors as Sian Reeves and Ben Daniels? Siân Reeves Siân Reeves (born Siân Rivers on May 9, 1966 in West Bromwich) is a British actress, most famous for playing the role of Sydney Henshall in the BBC drama "Cutting It", and for playing villain Sally Spode in "Emmerdale".
    What size population does the County Connection public transit in Concord, California service? County Connection The County Connection (officially, the Central Contra Costa Transit Authority, CCCTA) is a Concord-based public transit agency operating fixed-route bus and ADA paratransit (County Connection LINK) service in and around central Contra Costa County in the San Francisco Bay Area. Established in 1980 as a joint powers authority, CCCTA assumed control of public bus service within central Contra Costa first begun by Oakland-based AC Transit as it expanded into suburban Contra Costa County in the mid-1970s (especially after the opening of BART).
  • Loss: CachedMultipleNegativesBidirectionalRankingLoss with these parameters:
    {
        "temperature": 0.01,
        "similarity_fct": "cos_sim",
        "mini_batch_size": 128,
        "gather_across_devices": false
    }
    

Training Hyperparameters

Non-Default Hyperparameters

  • per_device_train_batch_size: 8192
  • per_device_eval_batch_size: 512
  • learning_rate: 0.0001
  • weight_decay: 0.01
  • num_train_epochs: 1
  • lr_scheduler_type: cosine
  • warmup_ratio: 0.1
  • seed: 12
  • bf16: True
  • dataloader_drop_last: True
  • dataloader_num_workers: 12
  • dataloader_prefetch_factor: 2
  • remove_unused_columns: False
  • optim: adamw_torch
  • batch_sampler: no_duplicates

All Hyperparameters

Click to expand
  • overwrite_output_dir: False
  • do_predict: False
  • eval_strategy: no
  • prediction_loss_only: True
  • per_device_train_batch_size: 8192
  • per_device_eval_batch_size: 512
  • 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: 0.0001
  • weight_decay: 0.01
  • adam_beta1: 0.9
  • adam_beta2: 0.999
  • adam_epsilon: 1e-08
  • max_grad_norm: 1.0
  • num_train_epochs: 1
  • max_steps: -1
  • lr_scheduler_type: cosine
  • 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: 12
  • data_seed: None
  • jit_mode_eval: 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: True
  • dataloader_num_workers: 12
  • dataloader_prefetch_factor: 2
  • past_index: -1
  • disable_tqdm: False
  • remove_unused_columns: False
  • 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}
  • parallelism_config: None
  • deepspeed: None
  • label_smoothing_factor: 0.0
  • optim: adamw_torch
  • optim_args: None
  • adafactor: False
  • group_by_length: False
  • length_column_name: length
  • project: huggingface
  • trackio_space_id: trackio
  • 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
  • hub_revision: None
  • 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: no
  • neftune_noise_alpha: None
  • optim_target_modules: None
  • batch_eval_metrics: False
  • eval_on_start: False
  • use_liger_kernel: False
  • liger_kernel_config: None
  • eval_use_gather_object: False
  • average_tokens_across_devices: True
  • prompts: None
  • batch_sampler: no_duplicates
  • multi_dataset_batch_sampler: proportional
  • router_mapping: {}
  • learning_rate_mapping: {}

Training Logs

Epoch Step Training Loss NanoClimateFEVER_cosine_ndcg@10 NanoDBPedia_cosine_ndcg@10 NanoFEVER_cosine_ndcg@10 NanoFiQA2018_cosine_ndcg@10 NanoHotpotQA_cosine_ndcg@10 NanoMSMARCO_cosine_ndcg@10 NanoNFCorpus_cosine_ndcg@10 NanoNQ_cosine_ndcg@10 NanoQuoraRetrieval_cosine_ndcg@10 NanoSCIDOCS_cosine_ndcg@10 NanoArguAna_cosine_ndcg@10 NanoSciFact_cosine_ndcg@10 NanoTouche2020_cosine_ndcg@10 NanoBEIR_mean_cosine_ndcg@10
0.0190 10 11.3289 - - - - - - - - - - - - - -
0.0381 20 7.5743 - - - - - - - - - - - - - -
0.0571 30 5.4003 - - - - - - - - - - - - - -
0.0762 40 3.399 - - - - - - - - - - - - - -
0.0952 50 2.7399 - - - - - - - - - - - - - -
0.1143 60 2.415 - - - - - - - - - - - - - -
0.1333 70 2.3843 - - - - - - - - - - - - - -
0.1524 80 1.9827 - - - - - - - - - - - - - -
0.1714 90 1.8858 - - - - - - - - - - - - - -
0.1905 100 1.7143 - - - - - - - - - - - - - -
0.2095 110 2.0079 - - - - - - - - - - - - - -
0.2286 120 1.8461 - - - - - - - - - - - - - -
0.2476 130 1.7032 - - - - - - - - - - - - - -
0.2667 140 1.6531 - - - - - - - - - - - - - -
0.2857 150 1.9902 - - - - - - - - - - - - - -
0.3048 160 1.6245 - - - - - - - - - - - - - -
0.3238 170 1.685 - - - - - - - - - - - - - -
0.3429 180 1.657 - - - - - - - - - - - - - -
0.3619 190 1.8747 - - - - - - - - - - - - - -
0.3810 200 1.4671 - - - - - - - - - - - - - -
0.4 210 1.5957 - - - - - - - - - - - - - -
0.4190 220 1.5083 - - - - - - - - - - - - - -
0.4381 230 1.5014 - - - - - - - - - - - - - -
0.4571 240 1.4548 - - - - - - - - - - - - - -
0.4762 250 1.5598 - - - - - - - - - - - - - -
0.4952 260 1.3879 - - - - - - - - - - - - - -
0.5143 270 1.5633 - - - - - - - - - - - - - -
0.5333 280 1.5092 - - - - - - - - - - - - - -
0.5524 290 1.4434 - - - - - - - - - - - - - -
0.5714 300 1.5024 - - - - - - - - - - - - - -
0.5905 310 1.511 - - - - - - - - - - - - - -
0.6095 320 1.4404 - - - - - - - - - - - - - -
0.6286 330 1.6083 - - - - - - - - - - - - - -
0.6476 340 1.4197 - - - - - - - - - - - - - -
0.6667 350 1.5548 - - - - - - - - - - - - - -
0.6857 360 1.5642 - - - - - - - - - - - - - -
0.7048 370 1.4709 - - - - - - - - - - - - - -
0.7238 380 1.482 - - - - - - - - - - - - - -
0.7429 390 1.5472 - - - - - - - - - - - - - -
0.7619 400 1.4899 - - - - - - - - - - - - - -
0.7810 410 1.3321 - - - - - - - - - - - - - -
0.8 420 1.5174 - - - - - - - - - - - - - -
0.8190 430 1.3945 - - - - - - - - - - - - - -
0.8381 440 1.5877 - - - - - - - - - - - - - -
0.8571 450 1.3143 - - - - - - - - - - - - - -
0.8762 460 1.5343 - - - - - - - - - - - - - -
0.8952 470 1.4968 - - - - - - - - - - - - - -
0.9143 480 1.4361 - - - - - - - - - - - - - -
0.9333 490 1.4353 - - - - - - - - - - - - - -
0.9524 500 1.3146 - - - - - - - - - - - - - -
0.9714 510 1.3722 - - - - - - - - - - - - - -
0.9905 520 1.3098 - - - - - - - - - - - - - -
0 521 - 0.3204 0.5013 0.7971 0.4595 0.6496 0.5915 0.2981 0.6279 0.9387 0.3413 0.5898 0.6514 0.4762 0.5571

Framework Versions

  • Python: 3.11.14
  • Sentence Transformers: 5.3.0.dev0
  • Transformers: 4.57.1
  • PyTorch: 2.8.0+cu129
  • Accelerate: 1.12.0
  • Datasets: 4.4.1
  • Tokenizers: 0.22.1

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",
}