laitimes

Five pictures are chaotic: AI raw pictures sweep the junior design lion

author:AI Tech Review
Five pictures are chaotic: AI raw pictures sweep the junior design lion

AI Wensheng diagram is to train a neural network to automatically generate high-fidelity, high-resolution, and even high-fidelity images. Because a single sentence can create a picture, spoiled users gradually begin to act arbitrarily.

There are those who want to say a word to the AI and let the AI do all the work, and there are those who want to preemptively register the copyright of AI-generated content, but in the end, it is also humans who find shortcuts in different fields.

1

Derived from ChatGPT's DALL· E 3

  • Official website address: https://chat.openai.com
  • The top-of-the-line model comes with a fee of $20 per month
  • You need to visit a website outside of China

Thanks to ChatGPT's semantic understanding, DALL· E 3 is the lowest learning cost image generation tool on the market. Comprehension in almost all languages is supported, and with clear and detailed descriptions, you can generate a graph that meets your expectations:

Five pictures are chaotic: AI raw pictures sweep the junior design lion

After "Speaking People", it was DALL· E 3's drawing technology, although the output image itself is not currently "the highest quality work in the field of AI graphics". But in the junior game, it is basically unbeatable.

ChatGPT-based DALL· The E 3 technically represents the highest level of human-computer interaction, supports multiple language input to generate images, and is currently highly practical for the creation of English text. Just state the text you want to draw and you'll get good feedback.

Five pictures are chaotic: AI raw pictures sweep the junior design lion

2

最出圈的Midjourney

  • Official website address: https://www.midjourney.com/home
  • Must be used with Discord, a social chat app
  • Such a strong function, but it has not been packaged and landed until now
  • There is a fee, and packages are available for $10, $30, and $60 per month

Midjourney is currently the highest-rated tool in terms of performance, usability, and even business value. When it comes to generating images, Midjourney offers a wealth of models and features. Including the highest fidelity V6 on the market, the two-dimensional style Niji model.

Midjourney is also out of the market for its ability to produce images that are closest to the level of real photography. Let the public feel the ability of AI painting most intuitively.

Five pictures are chaotic: AI raw pictures sweep the junior design lion

In the process of using Midjourney to generate images, you can adjust the prompt (keyword) to control the content of the generated images, and further use the functions provided by Midjourney after the images are generated, including increasing the resolution and other ways to further improve the image quality.

Five pictures are chaotic: AI raw pictures sweep the junior design lion

At present, Midjourney is the most practical AI mapping tool in the world, but the operation threshold and direct payment cost are also relatively high.

3

Dreamina with the largest user base

  • Official website address: https://dreamina.jianying.com/ai-tool/home
  • Individuals are the easiest to show the attack of the work (Jianying has heard of it, and Douyin has used it)
  • It's currently free for a limited time

Dreamina has produced images that are rarely stunning in quality from the perspective of 2024. However, relying on the background of Jianying, Dreamina has attracted a large number of users with daily creative habits, which has greatly improved the creative collection.

The works generated by Dreamina can be recreated, including partial redrawing and AI expansion. These features require a lot of experimentation, and perhaps one or two users' ideas won't bring much work, but it is very likely that bringing together users and sharing them on the platform will allow users to spread more value among users that may be useful to another user.

Five pictures are chaotic: AI raw pictures sweep the junior design lion

4

Eye-catching Tiangong AI picture generation

  • Official website: https://www.tiangong.cn/chat/text_gen_image/004
  • It can understand a variety of semantics, such as the nine-square grid
  • It's completely free
  • The models are well raised, and most of the pictures are in line with the conventional aesthetic good-looking

Tiangong's AI biographs are based on the understanding of natural semantics, and can be stated according to their own ideas in style. Most of the styles on the market can now be generated by Tiangong AI images.

Five pictures are chaotic: AI raw pictures sweep the junior design lion

However, in Tiangong AI image generation, it does not support more modifications to the picture. And there are many ambiguities in the understanding of details, such as the inability to distinguish between "pure white background" and "snow white background". On the other hand, there are few ghost animals in the pictures generated by Tiangong, and most of the pictures have the bonus of "good-looking".

Five pictures are chaotic: AI raw pictures sweep the junior design lion

5

Stable Diffusion

  • Official website: https://www.stablediffusion.com (no need to visit)
  • The team is in danger of disbanding at the end of April 2024
  • It can be deployed on a local PC and requires a certain amount of software technical ability
  • The cost depends on how much you invest in the hardware

It is also the best tool for local deployment to liberate the AI creation seal.

Because it is deployed locally, the performance of Stable Diffusion is limited by the equipment used, and although the design boundaries are freer, the efficiency and cost are relatively low.

Five pictures are chaotic: AI raw pictures sweep the junior design lion

Compared to AI tools deployed on public networks, Stable Diffusion allows users to deploy on their own dedicated devices and remove unnecessary (or necessary) restrictions.

Five pictures are chaotic: AI raw pictures sweep the junior design lion

Fragmented thoughts: "Wensheng Tu" is an ideal that is partial and comprehensive.

Generating a satisfactory work with a Prompt is like a utopia of the idealist, as long as human beings are still pursuing aesthetics, then the best visual feast is inseparable from the creation of human beings, and there is always more beauty on top of beauty.

AI is already a great tool right now, and its value is unquestionable. But in terms of the value of the tool, the most important thing is still the human who uses it.

AI is successful for the time being, but in short, there is still work to be done.

The author of this article (wechat: warlial) subjective evaluation, real experience, bent on playing fun products well. Welcome to any product exchange that is still in the process of being conceived or has been implemented.

Without the authorization of "AI Technology Review", it is strictly forbidden to reprint it in any way on the webpage, forum, and community!

Please leave a message in the background of "AI Technology Review" to obtain authorization for reprinting on the official account, and you need to indicate the source and insert the business card of this official account when reprinting.

Read on