This really is "seeing in the future" and fun application. It really helps and makes it easy for us before going to barbershop. Now I can finally have some hair on my head and try different hairstyles😅😂
Indeed, haha! I agree this may not be useful to all of us! Still, I love seeing such "real-world applications" of GANs that can actually be useful compared to cartoonifiers and etc.!
is there an online app where we can use this .. i really would like to try it with some pictures to deside a new hair style... i have a bunch of haircuts i like on pinteres and also i have some of my own pics... how do i blendem? best regards
Would this application work towards “seeing the future” with images of land and water levels? I can see this working with the amount of weather tracking data to show or simulate conditions that would predict severity on natural weather disasters. Additionally, Google earth has an excellent database of images to map the world as we know it too. The images could help predicate costs associated with damages and even help with forecasting resources to respond to such events. The approach would be similar to how we would take photos of a car damage when submitting a claim with our car insurance.
Practice and understanding the theory! Read the papers carefully and start with pix2pix. I would also read many introductory articles explaining how and why it works! Let me know if you need help finding such resources.
@@WhatsAI I should have been more specific. Thanks for the response. How do you deal with non-convex optimization. Usually you want to escape saddle points and in the case of Gans you want to find them. It is so hard to get them to train right. You have optimization and stability problem in training which makes them particularly hard to handle. The images I get of range from Pollack to Picasso, and I want a (denis) Peterson.
@@WhatsAI Hey, Hi! I really never thought that you reply. Thankyou. Well, Did you checked whether it's working without any errors for hairstyle change?
I have had the time to implement personally from what I’ve seen it works if you follow the steps correctly. You can check the issues if you face any problem as many help each other in there
@@WhatsAI Finally, I have a last query. Does the hair always transfer upto shoulder level? I mean, If target image hair is long then do the same long hair will transfer OR just transfer upto only shoulder level on input? This is something i need to know, please.
Well, The model is not working. We need to run it 1000 times, atlast it throws an error. If it generates the result, then it takes minimum 15 minutes. Now always it generate result but it throw error most times. I tired waiting and waiting
References:
►Read the full article: www.louisbouchard.ai/barbershop/
►Peihao Zhu et al., (2021), Barbershop, arxiv.org/pdf/2106.01505.pdf
►Project link: zpdesu.github.io/Barbershop/
►Code: github.com/ZPdesu/Barbershop
I need this! I completely relate with the fear of changing hair cuts. This is so cool thanks for sharing!😊
Haha my pleasure, Gen! 😁 I hope they will release the code soon, so excited to try that!
This really is "seeing in the future" and fun application. It really helps and makes it easy for us before going to barbershop.
Now I can finally have some hair on my head and try different hairstyles😅😂
Indeed, haha! I agree this may not be useful to all of us!
Still, I love seeing such "real-world applications" of GANs that can actually be useful compared to cartoonifiers and etc.!
Yeah this is not that great but still helpful than cartoonify. GAN is really a game changer at least in some areas.
is there an online app where we can use this .. i really would like to try it with some pictures to deside a new hair style... i have a bunch of haircuts i like on pinteres and also i have some of my own pics... how do i blendem? best regards
Would this application work towards “seeing the future” with images of land and water levels? I can see this working with the amount of weather tracking data to show or simulate conditions that would predict severity on natural weather disasters. Additionally, Google earth has an excellent database of images to map the world as we know it too. The images could help predicate costs associated with damages and even help with forecasting resources to respond to such events.
The approach would be similar to how we would take photos of a car damage when submitting a claim with our car insurance.
Very Nice Presentation - Thank You Louis - J'adore !
Thank you very much, tout le plaisir est pour moi! 😁
Hello, Is it possible to get different AI models that can showcase different wigs?
So how do we get this app on our phone?
What's the app name?
як я можу спробувати цю программу?
Anyone made this work?
I'm very interested in this, is there any colab link？ I can give it a try
what is the secret to getting gans to work
Practice and understanding the theory! Read the papers carefully and start with pix2pix. I would also read many introductory articles explaining how and why it works! Let me know if you need help finding such resources.
@@WhatsAI I should have been more specific. Thanks for the response. How do you deal with non-convex optimization. Usually you want to escape saddle points and in the case of Gans you want to find them. It is so hard to get them to train right. You have optimization and stability problem in training which makes them particularly hard to handle. The images I get of range from Pollack to Picasso, and I want a (denis) Peterson.
Appreciate it.
im goingto make this new software product, i have other idea of implementing it. 30.10.2023.
What about doing this for videos?
github repo is not working
don't know how to clone
thank you
It's not free
cuz i also had had
problems related to
Just amazing!
Indeed Stephen! 😁
a tutorial on how to install would be appreciated
Sama boat.
Downloaded Anaconda.
Told to use the Environment file.
File didn't work and spammed errors.
This project is useless atm.
@@buchanpeter +1 been there, done that
It's a brilliant idea..
Can we design something for textiles/clothing b2b /b2c data prediction model?
Hi Louis, Does the code for hair transfer released?
Hi John! It has been! The link is in the video description.
@@WhatsAI Hey, Hi! I really never thought that you reply. Thankyou.
Well, Did you checked whether it's working without any errors for hairstyle change?
I have had the time to implement personally from what I’ve seen it works if you follow the steps correctly. You can check the issues if you face any problem as many help each other in there
@@WhatsAI Finally, I have a last query. Does the hair always transfer upto shoulder level? I mean, If target image hair is long then do the same long hair will transfer OR just transfer upto only shoulder level on input?
This is something i need to know, please.
Well, The model is not working. We need to run it 1000 times, atlast it throws an error. If it generates the result, then it takes minimum 15 minutes. Now always it generate result but it throw error most times. I tired waiting and waiting
This is cool