home.social

Search

1000 results for “Llama”

  1. michabbb @[email protected] ·

    🔍 #LlamaOCR introduces advanced document processing using #Llama32Vision technology

    • 📦 Simple #npm package requiring only 5 lines of code to implement advanced #OCR capabilities

    • 🔄 Converts complex documents including receipts and tables into structured markdown format

    • 🛠️ Built on #TogetherAI API with three model options: free tier, 11B, and 90B parameters for varying performance needs

    • 🗺️ Development roadmap includes support for single/multi-page #PDF processing and #JSON output formats

    #opensource #artificialintelligence #documentprocessing

    Try the demo: llamaocr.com
    Source: github.com/Nutlope/llama-ocr

    #llamaocr #llama32vision #npm #ocr #togetherai #pdf
  2. Reel Tubes @[email protected] ·

    Encántame este vídeo de Mancomún sobre Llamafile! Aprende a executar modelos de IA de xeito sinxelo no teu propio ordenador, con explicacións claras e ligazón ao artigo. Perfecto para quen quere probar IA local e software libre. Anímate a experimentar! #Llamafile #IA #InteligenciaArtificial #SoftwareLibre #Mancomun #Galego #PeerTube
    v.eurorede.com/videos/watch/22

    #llamafile #ia #inteligenciaartificial #softwarelibre #mancomun #galego
  3. Hacker News @[email protected] ·

    llamafile: Distribute and Run LLMs with a Single File

    github.com/mozilla-ai/llamafile

    #HackerNews #llamafile #LLMs #MozillaAI #AItools #MachineLearning #DistributeRun

    #hackernews #llamafile #llms #mozillaai #aitools #machinelearning
  4. Davide Eynard (+mala) @mala ·
    CW: AI, LLMs

    blog.mozilla.ai/llamafile-retu

    has been instrumental in my first projects relying on language models (I purposely left “large” aside, as the one I used for is tiny).

    While looking for support for more recent models I tried other applications, but I never felt “at home” with them as I did with llamafile. This is why I am so happy of this move: breathing new life into it is, for me, giving ppl a new chance to better understand (what ppl now call) “” while tinkering with it.

    #llamafile #byota #ai
  5. Open Risk @[email protected] ·

    @naturzukunft

    There aren't that many choices to manage exploding information complexity so something must happen.

    Even if it is not the semantic web as we knew it, it must solve similar problems (and hence look similar - as form follows function).

    Removing frictions is key imho. See e.g. how using complex LLM models is now down to just one #llamafile

    Semantic web tooling (at least the public / open source part) never reached a sweetspot for broad adoption, always somewhat aloof

    @rzeta0

    #llamafile
  6. Davide Eynard (+mala) @mala ·

    is a Mozilla tool that packages a language model in a single executable file that will run on most platforms. It is 100% local and has been optimized to run on slower hardware, from Raspberry Pis to my 8yo laptop. It is based on llama.cpp which supports a plethora of models, not just LLMs: I chose all-minilm because it’s tiny (50MB) and has open code, research papers, and datasets. github.com/Mozilla-Ocho/llamaf huggingface.co/sentence-transf

    #llamafile
  7. Matthew Bandy @[email protected] ·

    Seriously, #llamafile is amazing. On-device #llm with almost zero effort. Spectacular.

    github.com/Mozilla-Ocho/llamaf

    #llamafile #llm
  8. ⊥ᵒᵚ Cᵸᵎᶺᵋᶫ∸ᵒᵘ ☑️ @[email protected] ·

    I open sourced a thing !

    A super simple web GUI for @mozilla 's #llamafile

    Should pass the SO test.

    codeberg.org/falken/simple-gpt

    #ai #linux #chatgpt #llamacpp

    #llamacpp #linux #llamafile #ai #chatgpt
  9. Robert :ferris: :emacs: 🚀 @[email protected] ·

    Running #llamafile on #nixos and with a #nix flakes dev shell. It took me a while until I got a working setup that uses nvidia/cuda. In the interest of crowdsourced time saving I thought I post this.

    With kudos to arestless.rest/blog/llamafile- but the NVCC_APPEND_FLAGS no longer seems to work as intended.

    Copy/pastable version at github.com/rksm/llamafile-with

    #llamafile #nixos #nix
  10. Habr @[email protected] ·

    Как мы тестировали большие языковые модели для модерации отзывов

    В приложении «Магнит: акции и доставка» можно оставлять отзывы на товары. Отзывы модерируются: мы публикуем те, которые считаем полезными для других покупателей, — они должны описывать потребительские свойства товара. Отклоняем все остальные: как правило, это жалобы на ценники, сервис в магазине, условия хранения либо просто нерелевантные тексты. Отзывы с жалобами обрабатывают службы поддержки и сервиса. Рассказываем о том, как мы попробовали применять большие языковые модели, чтобы автоматизировать модерацию отзывов.

    habr.com/ru/companies/magnit/a

    #llm #magnit_tech #бям #chatgpt #yandexgpt #gigachat #llama

    #llama #gigachat #yandexgpt #chatgpt #бям #magnit_tech
  11. Dr. Fortyseven 🥃 █▓▒░ @[email protected] ·

    I had #Llama31 summarize an article and I screamed at it and it screamed back. Fair play, sir.

    #llm

    #llama31 #llm
  12. Dr. Fortyseven 🥃 █▓▒░ @[email protected] ·

    Wow, #LLama31 sure knows a lot about Homestar Runner.

    #llm

    #llama31 #llm
  13. patpro @[email protected] ·

    Suite de mes aventures avec #llama31 (dernière mise à jour datant de quelques jours), sur un PC Windows avec un GPU (pas certain qu’il soit utilisé, mais c’est plus rapide qu’avec la VM linux).
    Toujours mon carré de la racine carrée de 512. Accrochez-vous, ça pique :

    La racine carrée de 512 est
    √512 = √(8²) = 8
    Et le carré de 8 est
    8² = 64
    Donc, le carré de la racine carrée de 512 est 64.

    La seconde itération est encore plus stupide : il donne la bonne réponse immédiatement puis dérape :)

    #llama31
  14. Scripter :verified_flashing: @[email protected] ·

    Llama 3.1: Deshalb verschenkt Mark Zuckerberg seine künstliche Intelligenz | ZEIT ONLINE
    zeit.de/digital/2024-07/llama- #KI #KISprachmodell #Llama31

    #ki #kisprachmodell #llama31
  15. Jan Eggers @[email protected] ·

    Die Eckdaten laut dem "The Information"-Artikel: 7 Milliarden Dollar Trainings- und KI-Betriebskosten im Jahr, dazu 1,5 Milliarden Personalkosten - und nach Abzug der Einnahmen ein Loch von 5 Milliarden Dollar jährlich. Und: seit der Veröffentlichung von #Llama31-405B gestern Open-Source-Konkurrenz auf technisch vergleichbarem Niveau. Von Konkurrenten wie Anthropic (Geld von Google) ganz zu schweigen.

    #llama31
  16. Gaël Duval - /e/OS & Murena @[email protected] ·
    CW: Expect to be chocked

    🤡 Meta claims that LLaMA 3.1 is open source.

    So let's ask LLaMA 3.1 if it is open source ! 🤖💪

    #llama31 #opensource

    linkedin.com/feed/update/urn:l

    #llama31 #opensource
  17. Gaertan @[email protected] ·

    Another artfight attack! this time for shironya and their character Orpheus!

    #llama #anthro #furry #artfight #artfight2024 #seafoam #stardust #art #illustration #digital #krita #sfw #sfwart #sfwartwork #sfwfurry #portrait

    #llama #anthro #furry #artfight #artfight2024 #seafoam
  18. cos @cos ·

    @llamasoft_ox @gilesgoat i'm going to Orb gig in a couple of months when they are playing in Helsinki.

    #orb
  19. TheLuigiplayer @[email protected] ·

    🎅The Santa Llama is here in #LlamaLegends!
    🍬Event Shop: Get it for 1,500 Candy Canes!
    🎄Llama Bells: Event runs until Jan 5th.
    Let's spread some holiday cheer! 🎁

    theluigiplayer.lamaz.dev/llama-legends/

    #Gacha #IndieDev #free #Christmas #gaming #webgame #browsergame #event #xmas #cards #cardgames #cardgame #collect #collections #collector #collecting #collection #collectibles #fun #funny #santa #santaclaus

    #llamalegends #gacha #indiedev #free #christmas #gaming
  20. Saemon Zixel @[email protected] ·

    А llama.cpp достаточно легко и просто скомпилировалась в моей 32битной altlinux. Зависимостей мизер. Ничего не потребовалось доустанавливать, компилить. При этом работает стабильно, не ругается, не сегфолтиться.

    Тестил с Vikhr-Llama-3.2-1B-Q8_0.gguf, которая на 1,2ГБ и знает русский язык. Скорость "чтения" промпта 2 токена/сек. А скорость генерации ответа 1 токен/сек. Для вопросов "не к спеху" можно использовать, но качество ответа так себе.

    Замечу, что компьютер у меня старенький: Pentium D E6300 на 2,8Ггц, поддерживает максимум SSSE3 и работает с памятью DDR2 на 4ГБ. По этому, то, что есть уже радует меня)

    #llama #llamacpp #linux #vikhr

    #llama #llamacpp #linux #vikhr
  21. N-gated Hacker News @[email protected] ·

    Ah, the classic saga of the 🦙 Llama Community License, where everyone is either a rebel without a cause or just didn't read the fine print. 🤔 Who knew that using "Built with Llama" was the new form of rebellion? 💥 Welcome to the wild west of AI where compliance is optional, but confusion is mandatory. 🤷‍♂️
    notes.victor.earth/youre-proba #LlamaCommunityLicense #AIRebellion #ComplianceConfusion #TechSaga #WildWestAI #HackerNews #ngated

    #llamacommunitylicense #airebellion #complianceconfusion #techsaga #wildwestai #hackernews
  22. Neuronus Computing @[email protected] ·

    Meet Llama 3 and GPT-4 — two cutting-edge AI models built to elevate your experience.
    If you need fast, efficient responses, Llama 3 is your go-to. Prefer deep, accurate insights? GPT-4 delivers.
    From daily tasks to complex problem-solving, these tools adapt to your needs.⚡🧠

    Want to know which suits you best? Read our blog to explore more!👉

    neuronus.net/en/blog/meta-ais-

    #Llama3 #GPT4 #AIModels #AI #MachineLearning #AIComparison #SpeedVsAccuracy #AdvancedAI #AIForTasks #AIPower #FutureOfAI #Neuronus

    #llama3 #gpt4 #aimodels #ai #machinelearning #aicomparison
  23. Pyrzout :vm: @[email protected] ·

    Cómo saber si nos están grabando la llamada telefónica mientras hablamos blog.elhacker.net/2025/10/como #llamadas #android #grabar #ios

    #llamadas #android #grabar #ios
  24. HuffPost [ES] @[email protected] ·

    🛒 ■ Un estudio apunta la estafa más normalizada por los españoles y no tiene nada que ver con llamadas y mensajes fraudulentos ■ Si no con el sistema.
    huffingtonpost.es/life/consumo

    #llamadas #consumo #fraude #estafas #actualidad #mensajes

    #llamadas #consumo #fraude #estafas #actualidad #mensajes
  25. HuffPost [ES] @[email protected] ·

    💳 ■ Piden decir estas 5 palabras al descolgar una llamada comercial: te cuelgan al momento ■ Estas llamadas pueden resultar muy molestas, sobre todo porque son para ofrecer cosas que no interesan, o porque ocurren en horas inoportunas.
    huffingtonpost.es/life/consumo

    #llamadas #consumo

    #llamadas #consumo
  26. HuffPost [ES] @[email protected] ·

    🔥 ■ Dan aviso de lo que hay que hacer si te llama un número con prefijo 621 ■ Mejor prevenir.
    huffingtonpost.es/virales/dan-

    #llamadas #virales #telefonos #estafas

    #llamadas #virales #telefonos #estafas
  27. HuffPost [ES] @[email protected] ·

    🌐 ■ Una llamada interceptada pone de manifiesto cómo se está desvaneciendo la confianza de los ciudadanos rusos ■ Las versiones oficiales sobre la invasión de Ucrania cada vez convencen menos.
    huffingtonpost.es/global/una-l

    #llamadas #global #rusia #propaganda #guerraucrania

    #llamadas #global #rusia #propaganda #guerraucrania
  28. N-gated Hacker News @[email protected] ·

    TurboQuant model weight compression now graces #Llamacpp, but only if you speak fluent Metal! 🏋️‍♂️ Meanwhile, everyone else waits for TheTom to bless us with a #CUDA port, assuming he ever emerges from the GitHub labyrinth of Pull Request 45. How many engineers does it take to compress a llama? 🤔
    github.com/TheTom/llama-cpp-tu #TurboQuant #Metal #PullRequest #HackerNews #ngated

    #llamacpp #cuda #turboquant #metal #pullrequest #hackernews
  29. HuffPost [ES] @[email protected] ·

    🎯 ■ Mucho cuidado con esta nueva estafa telefónica: te llama un contacto cercano pero es IA ■ Engaños cada vez más sofisticados.
    huffingtonpost.es/virales/much

    #llamada #virales #movil #estafa #tiktok #inteligenciaartifical

    #llamada #virales #movil #estafa #tiktok #inteligenciaartifical
  30. AI Daily Post @[email protected] ·

    G42 just released NANDA 87B, an open-source Hindi-English model built on Llama-3.1 70B in partnership with MBZUAI and Cerebras. It promises powerful multilingual capabilities for the community. Dive into the details! #G42 #NANDA87B #Llama31 #OpenSource

    🔗 aidailypost.com/news/g42-unvei

    #llama31 #opensource #g42 #nanda87b
Next page