home.social

#stabledifusion — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #stabledifusion, aggregated by home.social.

fetched live
  1. #AI #stableDifusion #languagemodels #libre

    If there's one problem no one is solving, it's how to verify a digital asset as human made, but also how to prevent AI from understanding the asset at all.

    We're talking images, audio, video. There needs to be a way to create a piece of content that is verifiably human made and that cannot be replicated by AI.

  2. Weekend binge-watching "Learning #clickhouse with Mark Needham". Most recent addition -- Querying Parquet Metadata with ClickHouse (an example based on #huggingface and #stableDifusion datasets) youtube.com/watch?v=HzVCdUnlHq

  3. I will be speaking at @WeAreDevs World Congress! Are you using ? Or have fun playing with ? This time we will have a look at what potential problems that will come with those models. Get your ticket here: wearedevelopers.com/world-cong

  4. Looks pretty cool. Regarding the Jacket... the word "cyborg" was in the list... #stabledifusion #machinelearning #AIArtwork

  5. @jackyan
    Is that an extension of "AIs produce the darndest art"?

    (Im a big user of #stableDifusion , but noticing my feeds are becoming inundated with quantity, rather than quality AI images)

  6. #AI #GenerativeAI #StableDifusion #Japan: "After Takagi and his research partner Shinji Nishimoto built a simple model to “translate” brain activity into a readable format, Stable Diffusion was able to generate high-fidelity images that bore an uncanny resemblance to the originals.

    The AI could do this despite not being shown the pictures in advance or trained in any way to manufacture the results.

    “We really didn’t expect this kind of result,” Takagi said.

    Takagi stressed that the breakthrough does not, at this point, represent mind-reading – the AI can only produce images a person has viewed.

    “This is not mind-reading,” Takagi said. “Unfortunately there are many misunderstandings with our research.”

    “We can’t decode imaginations or dreams; we think this is too optimistic. But, of course, there is potential in the future.”"

    aljazeera.com/economy/2023/4/7

  7. #chatGPT #stableDifusion

    I'm no lawyer, but it looks like the Copyright office does not think the AI generate images have copyright protection, but the human written text of the graphic novel does. And you can't list the AI as a co-author. This probably means your chatGPT texts might as well be public domain & if you can sell it for $5.00, someone else can sell it for $0.05.

    copyright.gov/docs/zarya-of-th
    (found via google)

  8. @lzg Time Wizard is the name of a really good show that should have existed in 1995 (image courtesy of #Stabledifusion)

  9. 1/n Primera iteración funcional de mgm (model gallery maker) para WebUI de #stableDifusion funcionando.
    Se trata de generar una galería en local de los modelos descargados desde @HelloCivitai, en las que se guarden los datos de cada uno y sean accesibles inapp.
    Está siendo ->

  10. Ejemplo de #ControlNet en #stableDifusion (tirando de normal map.
    Se mantiene la posición del sujeto, se modifica todo lo que se quiera.
    ---
    RT @javilop
    Juguemos a un juego...

    He generado esta imagen usando #midjourney o #stablediffusion?

    ¡Intenta adivinar! Oh, espera... ¿no será una foto real? t.co/EQz1iqhW9G
    twitter.com/javilop/status/162

  11. Ante la velocidad que está alcanzado el despliegue de la inteligencia artificial no sabemos si dejarnos llevar por la fascinación o considerar seriamente su amenaza
    #ChatGPT #openai #stableDifusion #dalle #midjourney #GPT3

    collateralbits.net/perplejidad

  12. Lo del #outpainting con #stableDifusion
    Probatura con "Las Meninas" de Velázquez.
    Original / Expandida

  13. I have used #stableDifusion to generate t-shirt prints for my wife for Christmas. It was fun, but a but tiring experience, because I had to go through hundreds of generated images.

    Buying tshirt with human artist design from teepublic would be easier and probably better, but shipping from US = customs+3 week delivery, compared to 2 days local print is a dealbreaker.

    P.S. Could you guess keywords?

  14. The song is dedicated to ColD Fusion

    sigmoid.social/@LChoshen/10946

    BTW I created the above picture as there is still work to do on #stableDifusion #dalle2
    to create Bert as the statue of liberty, or even just Bert...

  15. CW: AI chat and images

    Asked #gptchat to describe what an AI looked like. It couldn't really so that.. But we chatted more and it wrote a story about an AI that needed to describe its own avatar to a blind girl called Lilly.

    I ran the gpt description of the fictional AI through #stableDifusion and got loads of drawings. Here's a couple of results. (the gpt prompt is in the alt Text).

    All quite predictable what it churns out... But still.....nagging feeling..I'm not sure about any of it!

  16. Testando uma parada aqui: Vão demitir os Devs ? O fim do emprego de ilustrador ! #stabledifusion ... youtube.com/clip/Ugkx8C8ND7i2A

  17. I asked #stabledifusion a picture of a CEO, president, boss, nurse, ... Unsurprinsigly it's as gendered as you would fear. Soon most supporting media will be AI-generated, and it's such a missed opportunity, and a low bar, to stop gendered historical stereotypes of professions.