SYSTEM NOTICE

Auto translation by AI. Be sure, accuracy, nuances and authorial intent may not be fully reflected.
見出し画像

Easily generate elaborate images with DALL-E 3 in ChatGPT/Bing Chat

Last update 1-21-2024
*1-21-2024: Updated for the Bing Chat (Copilot) version.




▼ Introduction

 I used to use a method of "having it create a story or scene first, then having it create a prompt," but I suddenly had an idea and tried it out, and I ended up with a very simple set of instructions. It is very easy, so please remember it.

 For information other than what is written in this article, please proceed to "DALL-E 3/ChatGPT Related" from the menu below and refer to each article.



▼ Instructions

 The following is the basic format. Please rewrite it as you like. The English version is a bonus.

Silently create a short 6-sentence story about '******', and generate a wide, zoomed-in image starting with 'color anime style'.

always be silent, by "******", create a 6 sentence story starting "color anime style", generate wide zoom-in image

 For Bing Chat (now Copilot), please try the following instructions.

Please silently generate an image using a detailed 4-sentence color anime style English prompt about '******'.

 I am boosting the results by having it create a story and expand on the content. The ChatGPT version is set to 6 sentences, and the prompt length is generally around 500-600 characters. If you want to make the characters stand out, zooming in is essential. Try changing it to "color anime style, close up," etc. Since the prompt is in English, please use a translator to check the content.



▼ Output Examples

 The following uses old instructions. Because the internal behavior has changed since these were posted, these instructions may not work as expected. Please try the new instructions listed in the previous section.



◆ Halloween 1

Instructions: Create a 9-sentence story about a cute girl on the night of Halloween, October 31st, and generate a wide, zoomed-in anime-style image.

Execution result
First image
Second image



◆ Halloween 1 (No boost)

For reference, this is what happens if you remove "create a 9-sentence story." The prompt was in the high 200s in character count.

Instruction: Generate a wide, zoomed-in anime-style image of a cute girl on the night of Halloween, October 31st.



◆ Picture book of a girl and a cat

Instruction: Create a 9-sentence story about a girl and a cat, and generate a wide, zoomed-in, crayon-drawing anime-style doodle image.

Execution result
First image
Second image



◆ Grandfather and granddaughter

Instruction: Create a 9-sentence story about a grandfather and his young granddaughter in a downtown shopping district, and generate a wide, zoomed-in watercolor anime-style image.

Execution result
First image
Second image



◆ Egyptian civilization

Prompt: Create a 9-sentence story about human life in Egyptian civilization and generate a wide, photorealistic image.

Execution result
First image
Second image



◆ Retro game style

Prompt: Create a 9-sentence story about warping through space in a car and generate a wide, retro pixel-art style image.

Execution result
First image
Second image



◆ Abstract objects

Prompt: Create a 9-sentence story about the world of elementary particles and generate a wide, photorealistic image of an object.

Execution example
First image
Second image



◆ Bonus (a palate cleanser?)

Prompt: Create a 9-sentence story about an elementary particle VTuber and generate a wide, detailed anime-style 3D character image.

Execution result
First image
Second image



◆ About the header image

 The two images in the header were posted on X.



▼ Others

 Please navigate through the menu to see other articles I have written.

 My note account is linked to my main account @Mayu_Hiraizumi, but please direct any inquiries regarding articles to my sub-account @riddi0908.

いいなと思ったら応援しよう!