home.social

#dalle — Public Fediverse posts

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

  1. I cannot believe #chatgpt (via #Dalle) actually made this for me, and with beautiful fingers, hands, and text. I might make my own version of this comic through actual handmade art to prove I'm a more devoted person than those icon grifters. (Alt-text is the #GPT's visual concept. #aislop #aipicture #stupid #comicstrip #JesusChrist #christianity #religion #comic #onepanel

  2. Как я обучил нейросеть генерировать постеры с Нивой

    В жизни каждого мужчины наступает момент когда он хочет купить Ниву. Вот в 33 года и мне захотелось приобрести данный автомобиль. А пока я копил деньги на долгожданную покупку, захотелось какого нибудь мерча с Нивой. Быстрый поиск по популярным маркетплейсам не дал результатов, которые бы порадовали моего внутренного эстета. Было решено сгенерировать изображения с помощью нейросети. Опробовал Dall‑E и Midjorney, Flux — все не то, нейросети никогда не видели нашей отечественной Нивы. Подумав что я же инженер, решил дообучить Flux на открытых фотографиях Нивы из интернета.

    habr.com/ru/articles/852954/

    #ии #ии_и_машинное_обучение #dalle #dalle #dalle2 #dalle_2 #dalle_mini

  3. Please indulge me as I give another example of what I mean by this progress.

    When I first tried DALL-E I thought a good challenge would be trying to recreate my wife's roleplaying character. She's written tons of material across numerous forum posts, and thus has a good idea of what her character looks like.

    So, why not try and bring her to life?

    The results were...not great.

    #DALLE #dalle2 #aiart #aiartcommunity #generativeart #stablediffusion

  4. Another encountered issue was that it was extremely difficult to force the generator to draw 7 dwarfs. Notoriously, 5 fingers of a hand are also difficult, but the larger the number the smaller hope to get it by chance. A third difficulty I found was the impossibility to generate impossible scenes, typical for fairy tales, such as #RedCap popping out happily from the wolf's belly or #Gretel shoving the witch into the oven. A quick comparison with #dallE showed that @openai's model handles somewhat better two of the above challenges (other than quantities). Yet, it struggles with unwanted #artefacts

  5. Why #DeepNeuralNetworks need #Logic:

    Nick Shea (#UCL/#Oxford) suggests

    (1) Generating novel stuff (e.g., #Dalle's art, #GPT's writing) is cool, but slow and inconsistent.

    (2) Just a handful of logical inferences can be used *across* loads of situations (e.g., #modusPonens works the same way every time).

    So (3) by #learning Logic, #DNNs would be able to recycle a few logical moves on a MASSIVE number of problems (rather than generate a novel solution from scratch for each one).

    #CompSci #AI

  6. It's been a metaphorical fifteen minutes and generative ai like #DallE and #ChatGPT are already having a greater impact on society than Blockchain & crypto ever did. Really happy to see the cultural zeitgeist move to tech that actually DOES something! Smart minds will be arguing for years whether it's for good or ill, but in the meanwhile this profile of a fiction author using #SudoWrite probably has some tips that could be useful for #TTRPG GM's with their #GMPrep

    theverge.com/23520625/chatgpt-