(≧▽≦)/ Hello to all new users!
This guide is created for those who would like to try image generation using the Tensor website. It is especially suitable for complete beginners who have never generated images or used this kind of website before.
If you already have experience with image generation, I recommend checking out my other articles instead.
This article is also available in Thai for new users who prefer reading in Thai.
You can message me privately on Discord at majesty.pop if you’d like the Thai version
บทความนี้มีเขียนเวอร์ชันภาษาไทย สำหรับผู้ใช้หน้าใหม่ที่ต้องการอ่านเป็นภาษาไทย สามารถทักมาทางดิสคอร์ดส่วนตัวได้ที่ majesty.pop นะคะ
1.Go to the Tensor.art website. (Well, you’re obviously already here if you’re reading my article, lol.)
2.Then look for the Post section.
3.Look for an image style you like from the posts in those categories.
4.Click to open that image.
5.Look for the Remix button.
6.Click it.
In this example, I’m using a post from my account as a sample.
7.After entering the Workspace, you will see various tools like in the image.
Red box = The model and LoRA used to create this image.
Green box = The prompt of the remixed image.
Purple box (Negative Prompt) = For things you don’t want in the image. For beginners who use Remix, this is usually already filled in from the original image, so you can just use it as is.
Yellow box = Click to generate the image.
Blue box = Scroll down and make sure to turn this option on, so your images won’t come out blurry or out of focus.
For those who want to change hair color or eye color, simply look at the original prompt and edit the parts you want to change.
masterpiece, best quality, ultra-detailed, 1girl, half-body, very long voluminous curly hair, pastel rainbow colors, purple eyes, fair skin, wearing flowing light fantasy dress, pastel colors, elegant pose, magical forest or garden background, warm cinematic lighting, charming expression, detailed hair and eyes, realistic proportions, feminine body
From this prompt, if you want to change the hair color or eye color, just delete the old colors and replace them with the colors you want.
Then click Remix, and that’s it — you’ll get your first usable image. ✨
And for beginners who want to delete everything and replace it with their own prompt?
If you want something basic and simple, you can just come to ChatGPT and ask for help.
Suppose you want an image of a girl wearing a cute, fluffy Santa outfit, with a Santa hat, black hair, and blue eyes (without fixing the hairstyle or pose just a test generation).
Open ChatGPT and use this command:
"Write a prompt for Tensor.art: a woman wearing a cute fluffy Santa outfit, with a Santa hat, black hair, and blue eyes. Start the prompt with 1woman, solo."
That’s all you’ll get a ready-to-use prompt for image generation.
That’s it now you have your very first image generated from your own prompt. ✨
There’s also a cool trick: you can let AI extract a prompt from an image too.
Just upload the image you want, and ask ChatGPT to analyze and recreate the prompt for you.
(It won’t be 100% the same, but it’s good enough to work with!)
For the image generated from the extracted prompt, it will come out like this.
But when you let AI recreate a prompt from an image, remember to edit things like hair color, eye color, or hairstyle yourself otherwise, it will stick to the example image.
