DALL·E mini Reviews and details

Social recommendations and mentions

• Ask HN: What is the most impactful thing you've ever built?

  Co-created Dall-E Mini (now named Craiyon, and I am not involved anymore). It was among the first text to language models created independently. And it was fully open source. It also got covered by New York Times in the article covering Dall-E 2 by Cade Metz. Links: * GitHub: https://github.com/borisdayma/dalle-mini * Hugging Face Demo: https://huggingface.co/spaces/flax-community/dalle-mini * NYT article:... - Source: Hacker News / over 1 year ago

• I saved the best for last, wifeys 💜

  It is a text-to-image AI generator, the one I used is called DALL·E mini. It’s kinda fun to play with, you can try it out here. Source: almost 2 years ago

• Looks more like Pink Guy than actual Pink Guy

  Https://huggingface.co/spaces/flax-community/dalle-mini you can use the generator here. Source: almost 2 years ago

• 3 kilkspoints to whowver gets the person right

  Watch the latest klik video https://huggingface.co/spaces/flax-community/dalle-mini. Source: almost 2 years ago

• What if Darth Vader was in AC

  Here you go, mi amigo. Note that there's a lot of traffic occurring on the website because of the damn memes right now. Source: almost 2 years ago

• ai_irl

  It and its predecessor, DALL-E, are not available to the general public as OpenAI is wary of its power being used maliciously, but some trusted users and researchers have been given access. The architecture outlining how it was made is freely published, allowing people to create and train their own versions (see the publicly available DALL-E mini), but few organizations have the immense levels of computation and... Source: almost 2 years ago

• DALL-E Mini AI-Generated Watercolor Paintings of "Aliens Landing Near St. Louis"

  Last night I was playing with some Artificial Intelligence tools I found online and thought the community would get a kick out of the results. The system is called DALL-E 2 Mini and is open to the public, while a version that generates high-quality images is kept private. You simply feed it a sentence based on what you want to generate, in this case, "Watercolor painting of aliens landing near St. Louis in the... Source: almost 2 years ago

• Does anyone knows free alternative to Dall-E 2?

  Have you tried the Dalle-Mini space on HuggingFace? It's using Dalle-Mega now which is in training still. Not exactly the same but it's nice. https://huggingface.co/spaces/flax-community/dalle-mini. Source: almost 2 years ago

• What steps are involved to ba able to use DALL•E Mega?

  More evidence that that web app is not Mini is that this older DALL-E Mini web app states that a newer model is used at the newer web app. Source: almost 2 years ago

• DALL-E Mega web app (not from OpenAI) is available, although training is only 18% done. This model has ~2.6 billion parameters. Example: "a rubik's cube".

  No. The web app at Hugging Face for the old DALL-E mini now directs users to the new web app, and states that the model used in the newer web app is better. Source: about 2 years ago

• Crazy new AI fully generates images given just a text description

  This is a low quality older version which you can try https://huggingface.co/spaces/flax-community/dalle-mini. Source: about 2 years ago

• March 4th - Free Draw Friday

  I just generate an AI landscape image at the DALL·E Mini website and then try to draw it in my art app. DALL·E Mini is HERE. Source: about 2 years ago

• Image Generation with Rudall-e mini

  Your experiences are probably typical for DALL-E mini. Note that DALL-E mini is different from ruDALL-E. In addition to diffusion models and VQGAN, some other text-to-image models you may wish to try are minDALL-E, Laionide, and OFA. Source: about 2 years ago

• Fun AI Stuff to play with on AOTS Vibe Check

  DALL-E Mini - like a more flexible version of DREAM (same underlying AI technique, ie DALL-E). Source: about 2 years ago

• All the AI created by OpenAI wants is a house. F

  Link here: https://huggingface.co/spaces/flax-community/dalle-mini. Source: over 2 years ago

• VQGAN + CLIP first time trying this

  There is a list of VQGAN+CLIP systems here. Included on that list are some non-Colab web apps such as hypnogram.xyz. I haven't yet revamped the main post linked to there, so here is a list from another person. There are also text-to-image systems that don't use CLIP such as CogView 2 and DALL-E mini. Source: over 2 years ago

• A text-to-image AI rendering of "Brian Tamaki in jail"

  Just tried DALL-E mini but it didn't have as many balls on the wall. (https://huggingface.co/spaces/flax-community/dalle-mini for those who want to see for themselves). Source: over 2 years ago

• Using AI software I created Art Deco styled digital paintings! More in comments

  Someone posted a demo for DALL-E (a portmanteau of Salvador Dali and Wall-e) in case anyone else wants to play around with generative networks like the one NightCafe uses for free https://huggingface.co/spaces/flax-community/dalle-mini. Source: over 2 years ago

• AI generated photo using 'California' as a keyword

  Every post like this Ive seen Ive been telling people to just use Dall-E Mini instead since its free and you get like 8 images per prompt. Source: over 2 years ago

• An image created by an AI with the keywords "An abandoned house"

  You can also use Dall-E Mini if you dont want to deal with the credits bullshit. And you get multiple images per prompt. Source: over 2 years ago

• An image I made using an AI. The phrase used was "An abandoned prison covered in fog"

  If you dont want to deal with that "credits" system check out Dall-e Mini: https://huggingface.co/spaces/flax-community/dalle-mini. Source: over 2 years ago