{
  "id": 320684,
  "title": "Dall-E Will Win",
  "url": "/competitions/gan-getting-started/discussion/320684",
  "author_name": "",
  "post_date": "2022-04-22T21:55:24.561951700Z",
  "votes": null,
  "comment_count": 6,
  "views": 0,
  "content": "<p>I didn't read much into this competition.  But,  I have a feeling that who ever implements Dall-E 2 here will win.</p>",
  "messages": [
    {
      "id": "1764828",
      "postDate": "04/22/2022 21:55:24",
      "content": "<p>I didn't read much into this competition.  But,  I have a feeling that who ever implements Dall-E 2 here will win.</p>",
      "rawMarkdown": "I didn't read much into this competition.  But,  I have a feeling that who ever implements Dall-E 2 here will win.",
      "votes": null
    },
    {
      "id": "1767161",
      "postDate": "04/25/2022 06:24:47",
      "content": "<p>I don't think Dall-E will win this competition. There are many other methods that could be more successful.</p>",
      "rawMarkdown": "I don't think Dall-E will win this competition. There are many other methods that could be more successful.",
      "votes": null
    },
    {
      "id": "1767750",
      "postDate": "04/25/2022 15:53:30",
      "content": "<p>You could be very well correct.  I cannot wait to see the results.</p>",
      "rawMarkdown": "You could be very well correct.  I cannot wait to see the results.",
      "votes": null
    },
    {
      "id": "1812946",
      "postDate": "06/06/2022 11:35:56",
      "content": "<p>Well, you are required to generate images from one given style. A style GAN is clearly superior to DALL-E in that task. For instance, generating images is something DALLE-2 sometimes struggles with, while there are specific style GANs that generate faces quite well.</p>",
      "rawMarkdown": "Well, you are required to generate images from one given style. A style GAN is clearly superior to DALL-E in that task. For instance, generating images is something DALLE-2 sometimes struggles with, while there are specific style GANs that generate faces quite well.",
      "votes": null
    },
    {
      "id": "1857295",
      "postDate": "07/16/2022 03:00:21",
      "content": "<p>I do agree with this. DALL-E is good to generate image from text but for other tasks I highly doubt it's capability. Thoughts on cycle GAN vs style GAN for this specific task ? </p>",
      "rawMarkdown": "I do agree with this. DALL-E is good to generate image from text but for other tasks I highly doubt it's capability. Thoughts on cycle GAN vs style GAN for this specific task ?",
      "votes": null
    },
    {
      "id": "2101970",
      "postDate": "01/16/2023 10:22:01",
      "content": "<p>I believe I read a research over internet that has compared the text2img generation of these models. DallE2 won. However it did not include Stable diffusion at the time. But sometimes GAN went a little discretionary in facial generation. And for the question you asked, I believe styleGAN2 will be the winner here (just prediction not tested xd)</p>",
      "rawMarkdown": "I believe I read a research over internet that has compared the text2img generation of these models. DallE2 won. However it did not include Stable diffusion at the time. But sometimes GAN went a little discretionary in facial generation. And for the question you asked, I believe styleGAN2 will be the winner here (just prediction not tested xd)",
      "votes": null
    },
    {
      "id": "2101971",
      "postDate": "01/16/2023 10:22:43",
      "content": "<p>Dall-E isn't the best in image generation business. Has anyone tried out comparison between previous architectures and Stable Diffusion?</p>",
      "rawMarkdown": "Dall-E isn't the best in image generation business. Has anyone tried out comparison between previous architectures and Stable Diffusion?",
      "votes": null
    }
  ],
  "comments": [
    {
      "id": 1767161,
      "author_name": "",
      "author_url": "",
      "post_date": "04/25/2022 06:24:47",
      "content": "<p>I don't think Dall-E will win this competition. There are many other methods that could be more successful.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1767750,
          "author_name": "",
          "author_url": "",
          "post_date": "04/25/2022 15:53:30",
          "content": "<p>You could be very well correct.  I cannot wait to see the results.</p>",
          "votes": null,
          "replies": []
        }
      ]
    },
    {
      "id": 1812946,
      "author_name": "josepc",
      "author_url": "",
      "post_date": "06/06/2022 11:35:56",
      "content": "<p>Well, you are required to generate images from one given style. A style GAN is clearly superior to DALL-E in that task. For instance, generating images is something DALLE-2 sometimes struggles with, while there are specific style GANs that generate faces quite well.</p>",
      "votes": null,
      "replies": [
        {
          "id": 1857295,
          "author_name": "himanshuwagh",
          "author_url": "",
          "post_date": "07/16/2022 03:00:21",
          "content": "<p>I do agree with this. DALL-E is good to generate image from text but for other tasks I highly doubt it's capability. Thoughts on cycle GAN vs style GAN for this specific task ? </p>",
          "votes": null,
          "replies": [
            {
              "id": 2101970,
              "author_name": "",
              "author_url": "",
              "post_date": "01/16/2023 10:22:01",
              "content": "<p>I believe I read a research over internet that has compared the text2img generation of these models. DallE2 won. However it did not include Stable diffusion at the time. But sometimes GAN went a little discretionary in facial generation. And for the question you asked, I believe styleGAN2 will be the winner here (just prediction not tested xd)</p>",
              "votes": null,
              "replies": []
            }
          ]
        }
      ]
    },
    {
      "id": 2101971,
      "author_name": "",
      "author_url": "",
      "post_date": "01/16/2023 10:22:43",
      "content": "<p>Dall-E isn't the best in image generation business. Has anyone tried out comparison between previous architectures and Stable Diffusion?</p>",
      "votes": null,
      "replies": []
    }
  ],
  "raw_markdown_by_id": {
    "1764828": "I didn't read much into this competition.  But,  I have a feeling that who ever implements Dall-E 2 here will win.",
    "1767161": "I don't think Dall-E will win this competition. There are many other methods that could be more successful.",
    "1767750": "You could be very well correct.  I cannot wait to see the results.",
    "1812946": "Well, you are required to generate images from one given style. A style GAN is clearly superior to DALL-E in that task. For instance, generating images is something DALLE-2 sometimes struggles with, while there are specific style GANs that generate faces quite well.",
    "1857295": "I do agree with this. DALL-E is good to generate image from text but for other tasks I highly doubt it's capability. Thoughts on cycle GAN vs style GAN for this specific task ?",
    "2101970": "I believe I read a research over internet that has compared the text2img generation of these models. DallE2 won. However it did not include Stable diffusion at the time. But sometimes GAN went a little discretionary in facial generation. And for the question you asked, I believe styleGAN2 will be the winner here (just prediction not tested xd)",
    "2101971": "Dall-E isn't the best in image generation business. Has anyone tried out comparison between previous architectures and Stable Diffusion?"
  },
  "source": "meta"
}