I noticed the original clothing had wide straps but the generated image had spaghetti straps. This is not something that would go overlooked and could not be used say for example if selling the clothing, as clothing design is very intricate and particular. You can see the original around the 2:08 mark and the final generated at 3:27. This is something I've noticed not just in yours, but in all AI videos that place clothing, it always alters the clothing in such a way as to not make it usable.
I appreciate the attention to detail. You’re right that AI generated images can sometimes change details, and this is a known limitation with current technology. With tools like fooocus AI, it is possible to generate multiple images to find one that closely matches the original design, but in most cases, this similarity reaches 90% and is not completely the same as the original photo. Most of the time, these tutorials are taught for virtual influencers, and I think this level of similarity with the original image is not bad, and by using the inpaint tool, the image can be more edited.
Glad it was helpful! You can change the whole scarf with the source image of your choice. Do the same as I showed for changing clothes and use proper source image. Don't mask the extra part as it can change your image.
very good tutorial but it looks openart updated the UI and I can't find those options from the video, I'm interesting in changing clothes from an image to the model, will be awesome if you can help
Hi. Could you assist me how to preserve the details of graphics on T-shirt? For example, I tried a tshirt with Jurassic park as text with graphic jurassic logo but the result is weird as unreadable text and graphic. How to get to some near to the generation? Thankyou
Hi. To get accurate details with fooocus, you can go to Advanced, Developer Debug Mode, Inpaint section, and change settings like 'Inpaint Denoising Strength' and sometimes the results may not be exactly the same. Test this tool and let me know if it works for you huggingface.co/spaces/yisol/IDM-VTON
For creating tutorials I prefer to use Google Colab because I want to know the limitations of working with this tool, so that I can help the audience more if needed. You can use any GPU based on what they said in the Minimal Requirement section.
Glad you liked it! In fooocus ai you need to mask the part you want to change it. If I find other ai tools that don't need mask area, I'll make a tutorial for it.
At 3 minutes and 50 seconds into your video, you add a "young woman hand" PROMPT. The bracelet is on the arm and not on the hand. I would appreciate it if you could explain the purpose of that PROMPT. I need your help for the following case. On a jacket you have the brand logo on the pocket. I want to delete the logo. I select the logo with the "eraser" and as a Prompt I write "nothing". In the image it generates, remove the logo and then place another logo. What can you recommend to solve it?
No matter. you can write arm or body. The purpose was to remove the extra part. What do you mean by choosing a logo with an eraser? You want to remove the logo on the jacket, so use the brush and mask the logo, select inpaint or outpaint method, and don't use styles, for prompt write 'jacket'. Write the prompt based on what you want to see in this area, in my example using arm was better but also 'hand' did my job, for you when you remove the logo you want to see the jacket so write it.
Running Fooocus typically requires more computing power than a standard Android device. According to the github page, if you have a powerful system, you can run it locally on Windows, Mac or Linux. You can also use it online on Google Colab that I showed you.
But GPU disconnecting again and again in only 3 to 4 minutes, when we perform heavy task! Like when we add multiple images in image prompt. Android phone
When you use Colab, do your work step by step. For example, first do face swap, then change the pose and expression, then edit the clothes and hair, and then improve the details. In Colab's free plan, you have limited daily usage per account, and you can't do heavy tasks at the same time.
All I say is EXCELLENT!👍Power of battery, It's AI Edge Mastery, Really I'm eagerly waiting for your next video❤. one more thing that I want to address and will grow our ASTER! family is you must do Octagonal research, cross niching, use some professional effects, do brandings in your videos. It help you to earn and invest in your channel. feel free to ask I have a lot of ideas and will update you in future time to time.💯💯
I noticed the original clothing had wide straps but the generated image had spaghetti straps. This is not something that would go overlooked and could not be used say for example if selling the clothing, as clothing design is very intricate and particular. You can see the original around the 2:08 mark and the final generated at 3:27. This is something I've noticed not just in yours, but in all AI videos that place clothing, it always alters the clothing in such a way as to not make it usable.
I appreciate the attention to detail. You’re right that AI generated images can sometimes change details, and this is a known limitation with current technology. With tools like fooocus AI, it is possible to generate multiple images to find one that closely matches the original design, but in most cases, this similarity reaches 90% and is not completely the same as the original photo. Most of the time, these tutorials are taught for virtual influencers, and I think this level of similarity with the original image is not bad, and by using the inpaint tool, the image can be more edited.
Thanks for the tips. Can I change the model’s scarf with a particular design?
Glad it was helpful! You can change the whole scarf with the source image of your choice. Do the same as I showed for changing clothes and use proper source image. Don't mask the extra part as it can change your image.
can you create a workflow in comfyui as well
In this channel, I have not started the training for working with ComfyUI. Maybe I will do this in the future.
very good tutorial but it looks openart updated the UI and I can't find those options from the video, I'm interesting in changing clothes from an image to the model, will be awesome if you can help
Thanks! Which options can't you find? Let me know so I can help you.
Hi. Could you assist me how to preserve the details of graphics on T-shirt? For example, I tried a tshirt with Jurassic park as text with graphic jurassic logo but the result is weird as unreadable text and graphic. How to get to some near to the generation? Thankyou
Hi. To get accurate details with fooocus, you can go to Advanced, Developer Debug Mode, Inpaint section, and change settings like 'Inpaint Denoising Strength' and sometimes the results may not be exactly the same.
Test this tool and let me know if it works for you
huggingface.co/spaces/yisol/IDM-VTON
what is your hardware setup? especially nvidia GPU used ?
For creating tutorials I prefer to use Google Colab because I want to know the limitations of working with this tool, so that I can help the audience more if needed.
You can use any GPU based on what they said in the Minimal Requirement section.
Make a video if you can the same dress of the same model with only the pose style changed
Thanks for the video, How can I get the same dress with the same pattern without using the mask?
Glad you liked it! In fooocus ai you need to mask the part you want to change it. If I find other ai tools that don't need mask area, I'll make a tutorial for it.
I’d use DeFooocus instead as it have a built in background removal tool so you won’t have to leave the app to do that part.
Thanks for your feedback! We will introduce and use DeFooocus In future tutorials.
Your explanation is very good and straight to the point 👍🏻👏🏻❣️
Glad you liked it! Thanks for your kind feedback! 😊
Thank you that is a huge help 👍
Glad it helped! Thanks for your feedback! 😊
What a great video, very well explain it!
Glad you liked it! 😊
At 3 minutes and 50 seconds into your video, you add a "young woman hand" PROMPT. The bracelet is on the arm and not on the hand. I would appreciate it if you could explain the purpose of that PROMPT.
I need your help for the following case. On a jacket you have the brand logo on the pocket. I want to delete the logo. I select the logo with the "eraser" and as a Prompt I write "nothing". In the image it generates, remove the logo and then place another logo. What can you recommend to solve it?
No matter. you can write arm or body. The purpose was to remove the extra part.
What do you mean by choosing a logo with an eraser? You want to remove the logo on the jacket, so use the brush and mask the logo, select inpaint or outpaint method, and don't use styles, for prompt write 'jacket'. Write the prompt based on what you want to see in this area, in my example using arm was better but also 'hand' did my job, for you when you remove the logo you want to see the jacket so write it.
@@AIEdgeMastery Thanks. This tip helping me.
You're welcome. Good luck with that.
Extraordinary explanation
Thanks! Glad you liked it. 😊
since 32 images can be generated at once, how can we have multiple colors on same outfit put on at one go?
You can't do that
can't reproduce fabric texture when mixing
Fooocus depends a lot on how you set up prompt and settings. Have you tried several times?
@@AIEdgeMastery I have tried many times - the texture of the fabric is distorted and becomes different from the donor picture
Can i use it on Android?
Running Fooocus typically requires more computing power than a standard Android device.
According to the github page, if you have a powerful system, you can run it locally on Windows, Mac or Linux.
You can also use it online on Google Colab that I showed you.
I use Focus.
It also changes the model's body. What should I do to keep the same body and only change the clothes?
Just do as I did in this tutorial and nothing will change.
👍💓!!
👍👍
But GPU disconnecting again and again in only 3 to 4 minutes, when we perform heavy task! Like when we add multiple images in image prompt. Android phone
When you use Colab, do your work step by step. For example, first do face swap, then change the pose and expression, then edit the clothes and hair, and then improve the details.
In Colab's free plan, you have limited daily usage per account, and you can't do heavy tasks at the same time.
All I say is EXCELLENT!👍Power of battery, It's AI Edge Mastery, Really I'm eagerly waiting for your next video❤. one more thing that I want to address and will grow our ASTER! family is you must do Octagonal research, cross niching, use some professional effects, do brandings in your videos. It help you to earn and invest in your channel. feel free to ask I have a lot of ideas and will update you in future time to time.💯💯
Thanks for your guidance. I will try to improve the channel. 🙂