home.social

Search

1000 results for “recursive”

  1. C. @[email protected] ·

    #Python question...

    I recently ran into a surprise using #pathlib.Path - seems to be a sharp edge and I'm curious if others have run into it and what you thought.

    I wanted to find all #files/#dirs below the current #directory, and thought a #recursive #glob would do it - `Path.cwd().glob("**")`. I was surprised that this returns all directories but not files. "**/*" is needed to find the files as well.

    I'm not a big user of recursive globs, so maybe this is expected #behaviour?

    #python #pathlib #files #directory #recursive #glob
  2. Foucault_this @[email protected] ·

    Reading about the evaporation of #cholombiano #music #subculture and #subaltern #hair I bumped into a survey about what #home means to me and requesting my definition.

    Here's mine:
    Home is a node where you star from and return to recursively stitching experience back into home's mosaic hearth and wrap around yourself and your loved ones to dream before heading out to see the world again as if for the first time.

    What's your definition?
    #cozy #home #whatdoeshomemeantoyou #whatishome

    #cholombiano #music #subculture #subaltern #hair #home
  3. 📡 RightToPrivacy & Tech Tips @RTP ·

    :linux: LINUX TIP:

    Running Into "Permission Denied"?

    💡 You can easily change ownership of all files / directories (recursively) by running:

    sudo chown -R user *

    From inside the directory where you would like all files / directories changed to 'user' ownership.

    The '-R' flag makes chown perform recursively.

    The '*' matches / symbolizes all

    Try this next time you need to change ownership

    #linux #tips #chown #commandline #shell #bash
  4. Frédéric Jacobs @[email protected] ·

    Great highlight [1] by @QuantaMagazine on the work done [2] by Keegan Ryan and Nadia Heninger on improving the efficiency of the LLL algorithm using multiple techniques such as recursive structure and precision of numbers involved.
    Featuring @ducasleo

    [1]: quantamagazine.org/celebrated-
    [2]: iacr.org/cryptodb/data/paper.p
    #Lattices #cryptography #postquantum

    #lattices #cryptography #postquantum
  5. Bread80 @[email protected] ·

    I finally have the Flan wired up to the RGBToHDMI. Still some sparklies on screen but otherwise looking fantastic.

    And having a little explore of the BASIC. Recursive function definitions is cool. The wordiness isn’t so great though. Feels a bit like interpreted COBOL.

    #enterprise64

    #enterprise64
  6. Bread80 @bread80 ·

    I finally have the Flan wired up to the RGBToHDMI. Still some sparklies on screen but otherwise looking fantastic.

    And having a little explore of the BASIC. Recursive function definitions is cool. The wordiness isn’t so great though. Feels a bit like interpreted COBOL.

    #enterprise64
  7. Bread80 @[email protected] ·

    I finally have the Flan wired up to the RGBToHDMI. Still some sparklies on screen but otherwise looking fantastic.

    And having a little explore of the BASIC. Recursive function definitions is cool. The wordiness isn’t so great though. Feels a bit like interpreted COBOL.

    #enterprise64

    #enterprise64
  8. Bread80 @[email protected] ·

    I finally have the Flan wired up to the RGBToHDMI. Still some sparklies on screen but otherwise looking fantastic.

    And having a little explore of the BASIC. Recursive function definitions is cool. The wordiness isn’t so great though. Feels a bit like interpreted COBOL.

    #enterprise64

    #enterprise64
  9. Bread80 @[email protected] ·

    I finally have the Flan wired up to the RGBToHDMI. Still some sparklies on screen but otherwise looking fantastic.

    And having a little explore of the BASIC. Recursive function definitions is cool. The wordiness isn’t so great though. Feels a bit like interpreted COBOL.

    #enterprise64

    #enterprise64
  10. ƧƿѦςɛ♏ѦਹѤʞ @[email protected] ·

    Currently watching The Time Monster, episode 4, in which two TARDISes become recursively nested. One of greatest TV moments (for nerdy/mathsy kids in 1972) of all time. LOL
    #DoctorWho #TARDIS #plinge #thraskin #TOMTIT

    #doctorwho #tardis #plinge #thraskin #tomtit
  11. Jip J. Dekker @[email protected] ·

    I was scratching my head today about how to construct array literals `[1,2,3,…,n]` in #MiniZinc today. I was afraid that it would require some a bad recursive function, but it actually turns out to be quite easy once you remember the right builtin: `set2array(1..n)`

    #minizinc
  12. Habr @[email protected] ·

    Алгоритм поиска в глубину для процедурной генерации лабиринтов

    В этой статье я расскажу об алгоритме процедурной генерации лабиринтов методом поиска в глубину (Randomized depth-first search with recursive backtracking).

    habr.com/ru/articles/778202/

    #gamedev #maze #procedural_generation #depthfirstsearch

    #depthfirstsearch #procedural_generation #maze #gamedev
  13. Dave Walker 🎶 @[email protected] ·

    Ultimately, this is why search engines are doomed. It's really hard for a search engine to sniff this stuff out, and because they crawl this crap their overall indices are recursively polluted by it. #everythingisterrible

    futurism.com/sports-illustrate

    #everythingisterrible
  14. jackhutton 🇺🇸 @[email protected] ·

    Well, isn’t that just so interesting: “[..] there is a strong case that by definition a superintelligence would be fully impossible to control or contain. An “intelligence explosion” is the point at which an AI can improve itself again and again, recursively making itself better in ever faster and more effective ways. Here is the definitive uncontained and uncontainable technology. …

    #AI #theComingWave #tech #bookstodon

    #bookstodon #tech #thecomingwave #ai
  15. Colin Cogle 🔵 @[email protected] ·

    I've updated my MailPolicyExplainer #PowerShell module. Now, it can evaluate #SPF records recursively, counting how many #DNS lookups all of those "include" tokens consume. Does your SPF record return a PermError? Try the new `-Recurse` parameter and find out!

    Of course, it still checks the usual #email things: #DKIM, #DMARC, #BIMI, #DANE, #DNSSEC, #MTASTS, #NullMX records, and #IPv4 and #IPv6 reachability. Version 1.3.0 is now live in #PSGallery and on #GitHub.
    powershellgallery.com/packages

    #powershell #spf #dns #email #dkim #dmarc
  16. Max Resing @[email protected] ·

    Just recently, my colleagues at #NETSCOUT #ASERT published a new blog post on the importance of #DDoS defense mechanisms for both authoritative nameservers and recursive resolvers. It also sheds light on DDoS trends #DNS operators might want to pay attention to.

    netscout.com/blog/asert/the-po

    #netscout #asert #ddos #dns
  17. Max Resing @[email protected] ·

    Just recently, my colleagues at #NETSCOUT #ASERT published a new blog post on the importance of #DDoS defense mechanisms for both authoritative nameservers and recursive resolvers. It also sheds light on DDoS trends #DNS operators might want to pay attention to.

    netscout.com/blog/asert/the-po

    #netscout #asert #ddos #dns
  18. Max Resing @[email protected] ·

    Just recently, my colleagues at #NETSCOUT #ASERT published a new blog post on the importance of #DDoS defense mechanisms for both authoritative nameservers and recursive resolvers. It also sheds light on DDoS trends #DNS operators might want to pay attention to.

    netscout.com/blog/asert/the-po

    #netscout #asert #ddos #dns
  19. Max Resing @[email protected] ·

    Just recently, my colleagues at #NETSCOUT #ASERT published a new blog post on the importance of #DDoS defense mechanisms for both authoritative nameservers and recursive resolvers. It also sheds light on DDoS trends #DNS operators might want to pay attention to.

    netscout.com/blog/asert/the-po

    #dns #ddos #asert #netscout
  20. Tero Keski-Valkama @[email protected] ·

    We fine-tune custom #LLMs for two main reasons:
    - To conserve precious context tokens, and
    - To introduce the #LLM to some new knowledge or skill that wasn't available for its generalist training set.

    Fine-tuning is not a solution for utilizing personal or confidential data! The fine-tuned models will leak this information.

    So let's assume we aren't working with private data.

    In general, because of transfer learning, it would in principle make more sense to incorporate the new knowledge into the base model corpus, because that tends to create better models. But still, even if the generalist model knows your data and the task, if you're going to put that generalist model into a component of your larger system where it will always perform the same task, it makes sense to fine-tune it for this task only rather than to feed the same prompt prefix to it for every inference round.

    Now with data-centric #AI it might even be that the data you want to use doesn't meet the high quality standards large generalist models require. Perhaps in these cases it might make sense to let a chatbot rewrite your specialist corpus into a higher quality form, even if you're not aiming to incorporate your data into generalist corpuses.

    There is a new use case emerging though, #RecursiveSelfImprovement. I believe we can do this in a synergistic generalist fashion as well, but curiously it's now something even smaller organizations can do for specialized tasks by fine-tuning.

    Much like #alignment, it went from niche philosophical topic into standard engineering practices overnight.

    Recursive self-improvement is done by #DataCentricAI principles where a fine-tuned task is trained by examples, but those examples are generated and filtered recursively by the LLM. In principle the model is fine-tuned in rounds, using e.g. #DPO. In a round, the model is first fine-tuned with the existing good data. Then it's asked to generate new variations for those examples. Then its asked to rank pairs of training data examples and the worse ones are filtered out. Then the resulting dataset now has more task examples but of better quality than before. This is again used for fine-tuning and the cycle starts again.

    As this isn't human-imitative, the chatbots can exceed human parity.

    It requires a bit of nuance though. There is not only one task this specialist bot is taught but a set:
    1. Generate variations of tasks (including this task itself).
    2. Rank pairs of task performances (including ranking task).
    3. Perform the task proper.

    #llms #llm #ai #recursiveselfimprovement #alignment #datacentricai
  21. Tero Keski-Valkama @[email protected] ·

    We fine-tune custom #LLMs for two main reasons:
    - To conserve precious context tokens, and
    - To introduce the #LLM to some new knowledge or skill that wasn't available for its generalist training set.

    Fine-tuning is not a solution for utilizing personal or confidential data! The fine-tuned models will leak this information.

    So let's assume we aren't working with private data.

    In general, because of transfer learning, it would in principle make more sense to incorporate the new knowledge into the base model corpus, because that tends to create better models. But still, even if the generalist model knows your data and the task, if you're going to put that generalist model into a component of your larger system where it will always perform the same task, it makes sense to fine-tune it for this task only rather than to feed the same prompt prefix to it for every inference round.

    Now with data-centric #AI it might even be that the data you want to use doesn't meet the high quality standards large generalist models require. Perhaps in these cases it might make sense to let a chatbot rewrite your specialist corpus into a higher quality form, even if you're not aiming to incorporate your data into generalist corpuses.

    There is a new use case emerging though, #RecursiveSelfImprovement. I believe we can do this in a synergistic generalist fashion as well, but curiously it's now something even smaller organizations can do for specialized tasks by fine-tuning.

    Much like #alignment, it went from niche philosophical topic into standard engineering practices overnight.

    Recursive self-improvement is done by #DataCentricAI principles where a fine-tuned task is trained by examples, but those examples are generated and filtered recursively by the LLM. In principle the model is fine-tuned in rounds, using e.g. #DPO. In a round, the model is first fine-tuned with the existing good data. Then it's asked to generate new variations for those examples. Then its asked to rank pairs of training data examples and the worse ones are filtered out. Then the resulting dataset now has more task examples but of better quality than before. This is again used for fine-tuning and the cycle starts again.

    As this isn't human-imitative, the chatbots can exceed human parity.

    It requires a bit of nuance though. There is not only one task this specialist bot is taught but a set:
    1. Generate variations of tasks (including this task itself).
    2. Rank pairs of task performances (including ranking task).
    3. Perform the task proper.

    #llms #llm #ai #recursiveselfimprovement #alignment #datacentricai
  22. Tero Keski-Valkama @[email protected] ·

    We fine-tune custom #LLMs for two main reasons:
    - To conserve precious context tokens, and
    - To introduce the #LLM to some new knowledge or skill that wasn't available for its generalist training set.

    Fine-tuning is not a solution for utilizing personal or confidential data! The fine-tuned models will leak this information.

    So let's assume we aren't working with private data.

    In general, because of transfer learning, it would in principle make more sense to incorporate the new knowledge into the base model corpus, because that tends to create better models. But still, even if the generalist model knows your data and the task, if you're going to put that generalist model into a component of your larger system where it will always perform the same task, it makes sense to fine-tune it for this task only rather than to feed the same prompt prefix to it for every inference round.

    Now with data-centric #AI it might even be that the data you want to use doesn't meet the high quality standards large generalist models require. Perhaps in these cases it might make sense to let a chatbot rewrite your specialist corpus into a higher quality form, even if you're not aiming to incorporate your data into generalist corpuses.

    There is a new use case emerging though, #RecursiveSelfImprovement. I believe we can do this in a synergistic generalist fashion as well, but curiously it's now something even smaller organizations can do for specialized tasks by fine-tuning.

    Much like #alignment, it went from niche philosophical topic into standard engineering practices overnight.

    Recursive self-improvement is done by #DataCentricAI principles where a fine-tuned task is trained by examples, but those examples are generated and filtered recursively by the LLM. In principle the model is fine-tuned in rounds, using e.g. #DPO. In a round, the model is first fine-tuned with the existing good data. Then it's asked to generate new variations for those examples. Then its asked to rank pairs of training data examples and the worse ones are filtered out. Then the resulting dataset now has more task examples but of better quality than before. This is again used for fine-tuning and the cycle starts again.

    As this isn't human-imitative, the chatbots can exceed human parity.

    It requires a bit of nuance though. There is not only one task this specialist bot is taught but a set:
    1. Generate variations of tasks (including this task itself).
    2. Rank pairs of task performances (including ranking task).
    3. Perform the task proper.

    #llms #llm #ai #recursiveselfimprovement #alignment #datacentricai
  23. Tero Keski-Valkama @[email protected] ·

    We fine-tune custom #LLMs for two main reasons:
    - To conserve precious context tokens, and
    - To introduce the #LLM to some new knowledge or skill that wasn't available for its generalist training set.

    Fine-tuning is not a solution for utilizing personal or confidential data! The fine-tuned models will leak this information.

    So let's assume we aren't working with private data.

    In general, because of transfer learning, it would in principle make more sense to incorporate the new knowledge into the base model corpus, because that tends to create better models. But still, even if the generalist model knows your data and the task, if you're going to put that generalist model into a component of your larger system where it will always perform the same task, it makes sense to fine-tune it for this task only rather than to feed the same prompt prefix to it for every inference round.

    Now with data-centric #AI it might even be that the data you want to use doesn't meet the high quality standards large generalist models require. Perhaps in these cases it might make sense to let a chatbot rewrite your specialist corpus into a higher quality form, even if you're not aiming to incorporate your data into generalist corpuses.

    There is a new use case emerging though, #RecursiveSelfImprovement. I believe we can do this in a synergistic generalist fashion as well, but curiously it's now something even smaller organizations can do for specialized tasks by fine-tuning.

    Much like #alignment, it went from niche philosophical topic into standard engineering practices overnight.

    Recursive self-improvement is done by #DataCentricAI principles where a fine-tuned task is trained by examples, but those examples are generated and filtered recursively by the LLM. In principle the model is fine-tuned in rounds, using e.g. #DPO. In a round, the model is first fine-tuned with the existing good data. Then it's asked to generate new variations for those examples. Then its asked to rank pairs of training data examples and the worse ones are filtered out. Then the resulting dataset now has more task examples but of better quality than before. This is again used for fine-tuning and the cycle starts again.

    As this isn't human-imitative, the chatbots can exceed human parity.

    It requires a bit of nuance though. There is not only one task this specialist bot is taught but a set:
    1. Generate variations of tasks (including this task itself).
    2. Rank pairs of task performances (including ranking task).
    3. Perform the task proper.

    #dpo #datacentricai #alignment #recursiveselfimprovement #ai #llm
  24. Tero Keski-Valkama @[email protected] ·

    We fine-tune custom #LLMs for two main reasons:
    - To conserve precious context tokens, and
    - To introduce the #LLM to some new knowledge or skill that wasn't available for its generalist training set.

    Fine-tuning is not a solution for utilizing personal or confidential data! The fine-tuned models will leak this information.

    So let's assume we aren't working with private data.

    In general, because of transfer learning, it would in principle make more sense to incorporate the new knowledge into the base model corpus, because that tends to create better models. But still, even if the generalist model knows your data and the task, if you're going to put that generalist model into a component of your larger system where it will always perform the same task, it makes sense to fine-tune it for this task only rather than to feed the same prompt prefix to it for every inference round.

    Now with data-centric #AI it might even be that the data you want to use doesn't meet the high quality standards large generalist models require. Perhaps in these cases it might make sense to let a chatbot rewrite your specialist corpus into a higher quality form, even if you're not aiming to incorporate your data into generalist corpuses.

    There is a new use case emerging though, #RecursiveSelfImprovement. I believe we can do this in a synergistic generalist fashion as well, but curiously it's now something even smaller organizations can do for specialized tasks by fine-tuning.

    Much like #alignment, it went from niche philosophical topic into standard engineering practices overnight.

    Recursive self-improvement is done by #DataCentricAI principles where a fine-tuned task is trained by examples, but those examples are generated and filtered recursively by the LLM. In principle the model is fine-tuned in rounds, using e.g. #DPO. In a round, the model is first fine-tuned with the existing good data. Then it's asked to generate new variations for those examples. Then its asked to rank pairs of training data examples and the worse ones are filtered out. Then the resulting dataset now has more task examples but of better quality than before. This is again used for fine-tuning and the cycle starts again.

    As this isn't human-imitative, the chatbots can exceed human parity.

    It requires a bit of nuance though. There is not only one task this specialist bot is taught but a set:
    1. Generate variations of tasks (including this task itself).
    2. Rank pairs of task performances (including ranking task).
    3. Perform the task proper.

    #llms #llm #ai #recursiveselfimprovement #alignment #datacentricai
  25. Crypto News @[email protected] ·

    Bitcoin Witnesses Second-Highest Daily Ordinal Inscription Count as Inscriptions Near 30 Million - While Bitcoin-based nonfungible token (NFT) sales have declined, and the network h... - news.bitcoin.com/bitcoin-witne #ordinalinscriptions #cryptoslam.io #duneanalytics #transactions #uniqueusers #$30million #blockchain #dailycount #recursives #magiceden #milestone #nftsales #bitcoin #decline #buyers

    #buyers #decline #bitcoin #nftsales #milestone #magiceden
  26. Crypto News @[email protected] ·

    Bitcoin Witnesses Second-Highest Daily Ordinal Inscription Count as Inscriptions Near 30 Million - While Bitcoin-based nonfungible token (NFT) sales have declined, and the network h... - news.bitcoin.com/bitcoin-witne #ordinalinscriptions #cryptoslam.io #duneanalytics #transactions #uniqueusers #$30million #blockchain #dailycount #recursives #magiceden #milestone #nftsales #bitcoin #decline #buyers

    #buyers #decline #bitcoin #nftsales #milestone #magiceden
  27. Crypto News @[email protected] ·

    Bitcoin Witnesses Second-Highest Daily Ordinal Inscription Count as Inscriptions Near 30 Million - While Bitcoin-based nonfungible token (NFT) sales have declined, and the network h... - news.bitcoin.com/bitcoin-witne #ordinalinscriptions #cryptoslam.io #duneanalytics #transactions #uniqueusers #$30million #blockchain #dailycount #recursives #magiceden #milestone #nftsales #bitcoin #decline #buyers

    #buyers #decline #bitcoin #nftsales #milestone #magiceden
  28. Crypto News @[email protected] ·

    Bitcoin Witnesses Second-Highest Daily Ordinal Inscription Count as Inscriptions Near 30 Million - While Bitcoin-based nonfungible token (NFT) sales have declined, and the network h... - news.bitcoin.com/bitcoin-witne #ordinalinscriptions #cryptoslam.io #duneanalytics #transactions #uniqueusers #$30million #blockchain #dailycount #recursives #magiceden #milestone #nftsales #bitcoin #decline #buyers

    #ordinalinscriptions #cryptoslam #duneanalytics #transactions #uniqueusers #blockchain
  29. Amy Zhang @[email protected] ·

    Try out this new demo of a tool for expandable, explorable, recursive summary artifacts using LLMs from Semantic Scholar! 🤩 You can select any text span to expand at that spot & even dive into the original source.

    exp-sum.apps.allenai.org It's currently loaded up with #VLDB2023 papers!

    #vldb2023
  30. Amy Zhang @[email protected] ·

    Try out this new demo of a tool for expandable, explorable, recursive summary artifacts using LLMs from Semantic Scholar! 🤩 You can select any text span to expand at that spot & even dive into the original source.

    exp-sum.apps.allenai.org It's currently loaded up with #VLDB2023 papers!

    #vldb2023
Next page