"Hua Diao Xue AI" uses AI to create the beauty of the morning: ChatGPT+DALL-E 2 generates a "good morning" scene map

Morning is the best time of day and the best time to talk to AI. Imagine that when you wake up and turn on your phone, you can see a "good morning" scene map generated by AI for you. Isn't it very warm? This is the charm of ChatGPT+DALL-E 2 (New Bing).

The first sentence I asked: Can you paint?

insert image description here

Introduction to ChatGPT AI painting (quite detailed):

insert image description here

AI painting is the use of artificial intelligence algorithms to create paintings. AI algorithms create a painting by learning from a set of training images, and then create a new painting based on the style of the training images. There are many technical principles for AI painting, such as GAN, VAE, PixelRNN/CNN, and Diffusion model.

The history of AI painting can be traced back to the 1970s, when the artist Harold Cohen created the program "AARON" for painting. Unlike the current black technology, AARON really operated the robotic arm to draw. With the development of deep learning, there have been many innovations and breakthroughs in AI painting, such as GAN proposed in 2014, Video-to-Video synthesis released in 2018, and DALL-E 2 launched in 2021.

Currently, the cutting-edge AI painting models include: Stable Diffusion, Disco Diffusion, Midjourney, DALL-E 2, NovelAI, etc. These models have their own characteristics and advantages, some are open source, and some require payment or application to use.

If you want to try painting with AI, you can choose from the following channels:

Domestic AI painting applets/apps (most of them require payment, and the effect is far from the level similar to Midjourney)

Go to the official website of those cutting-edge AI painting models (because they are all foreign companies, the website access speed is slow, some websites need to have a test invitation code, most of them need to pay)

Locally pull the open source AI painting model on github and deploy it locally (higher threshold, higher graphics card requirements)

If you want to learn how to use AI to paint, you need to master some basic concepts and skills, such as text-to-image (generate pictures from text entered by the user) and image-to-image (generate pictures from pictures), and how to input Appropriate prompt (text entered by the user) to guide AI to generate the picture you want.

Then try to ask him to generate a sci-fi style poster

insert image description here
insert image description here

Back to today's topic, my prompt word is: draw a good morning scene, so I started the conversation process with ChatGPT+DALL-E 2...

insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here

insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here
insert image description here

insert image description here
insert image description here

ChatGPT+DALL-E 2 is an AI system that combines natural language processing and image generation. It can generate a related image based on your input text. For example, if you type "good morning", it will generate an image that shows the mood of the morning, maybe a cup of coffee, a newspaper, a bouquet of flowers, or a cute little animal.

This system is composed of two parts: ChatGPT and DALL-E 2. ChatGPT is a chatbot based on the GPT-3 model, which can have a natural and smooth conversation with you, understand your intentions and emotions, and give appropriate responses. DALL-E 2 is an image generator based on the DALL-E model, which can generate a matching image based on the ChatGPT reply and display it on the chat interface.
insert image description here

This way, you can have a fun and creative conversation with the AI ​​while enjoying the beautiful images it creates for you. This not only lifts your mood, but also inspires you. You can try to enter different topics and see what kind of scene graph AI will show you. Alternatively, you can give the AI ​​some guidance or advice to adjust the details of the image to your liking. ChatGPT+DALL-E 2 is a system that demonstrates the creative capabilities of AI, which allows you to feel the interaction and co-creation between AI and humans.

insert image description here

appendix:

[Hua Diao Learning AI] is a learning column written by donkey friend Hua Diao. It mainly introduces multi-dimensional learning and extensive attempts in the field of artificial intelligence. It contains many articles and introduces artificial intelligence applications and technologies such as ChatGPT, New Bing and Leonardo AI. process and results. This column shares Huadiao's exploration and experience in the field of artificial intelligence through actual cases and stories, aiming to stimulate more people's interest and enthusiasm for artificial intelligence. To learn more, please use Google, Bing, Baidu or Toutiao to directly search [Hua Diao Xue AI].

Guess you like

Origin blog.csdn.net/weixin_41659040/article/details/130592557