Yeah I have been using this for months and the thing is that it still requires creativity and photoshop skills to pull of but yeah its super powerful and makes drafting ideas super quick
Thanks for this. QUESTION: How do you have multiple trained models in one shot? More specifically, is there a way to train several individual models, then bring them all together into one shot?
I would like to create portraits for my family members, can I make multiple image models in the same account in regenerate? Just by uploading their images and using different names?
Hi. Can I use a LoRA trained on Replicate with other accounts? In other words, can I train my LoRA with my face or a consistent character and share it for others to use with their accounts? If so, how does the process work? Thanks a lot
I got an error when training the Lora saying "Training failed. Error no file named config.json found in directory FLUX.1-dev/transformer." What do I do? HELP!
Great video! Thank you! I did the training - it completed, but I can't get it to give images. in the output it gives me an empty zip folder instead of images. I'm baffled. Although in my setup, I was never prompted for a trigger word. 🤷🏽♂
Hi any suggestions on why Replicate will not accept my Zip File. I tried your same naming format photo_James (1) and I noticed you used PNG's when there website suggested jpg's. I tried both and both files would not upload, nor did a drag to copy. TIA, James
Yes I wanted to keep this is simple to follow as possible and avoid technical terminology. Replicate stores the model within your account and you can use it on there or am I missing something?
What base model is this based on i.e. Juggertnaut, RealisticVision? Once model has been trained, does anyone know how to use downloaded weights in another stable diffusion platform like ComfyUI?
I tested the video you recommended earlier (you had seen this video before) and it didn't work. Now I did as you taught and it worked very well, thank you!
I put the cost into the video... cents. And no this is online as it runs on replicate gpus but yes with another workflow you could train a local model with you own images
What can i do if it doesnt detect the file i upload as an imput? already tried with jpeg and png files, then created the zip file to upload, but it keepds empty
I had the same problem. @mikeys4u comment above helped me. It's the browser. Firefox didn't work for me or for him. Chrome worked for him, I'm using Edge
absolutely love this, however, after I trained myself as my nickname Halligator, most of the images comes out as an alligator in stead of the trained images of me :) Any way to change the promt name? or do I have to train a new model?
Oh no sorry to hear that but then also that's kinda funny. You will have to train a new model. Good learning here for others: the name should be unambiguous.
I had the same problem. @mikeys4u comment above helped me. It's the browser. Firefox didn't work for me or for him. Chrome worked for him, I'm using Edge.
Done it. Its great. But it only ever generates me. So when i tell to generate a goofy putin image, it generates putin with my face. Even when i dont mention that myself... anyone with similar experiences ?
no. in the video he says it only createa images of you when you use the trigger word. in my case it always creates me. even if i don't use the trigger in my prompt.
IGOR YOU SHOULD HAVE POSTED IT YESTERDAY!!! Would have saved me lots of time with my recent thumbnail 😭😭😭 Anyhow, talk to you soon! (finishing up my BSc on Sep 9th, and then my brain is finally available for other stuff! 😀)
Amazing. I trained using 10 photos. And used the same prompt as you and the outputs were amazing. Thanks for $10 credit link. It worked.
I recommend you add "put all prompts into a copy paste box" and "describe each prompt in one concise casual sentence" in your custom gpt
That's a great tip thanks! Only downside is that you don't get to see the full prompt as it's all in one line.
My question now is if you're going to br using this for generating UA-cam thumbnails, how many times can you use it before you run out of credits?
Do other people have access to this trained model of yours or it stays private only to you?
I feel like this is how most youtubers will make thumbnails from now on. Just so easy and effective
Yeah I have been using this for months and the thing is that it still requires creativity and photoshop skills to pull of but yeah its super powerful and makes drafting ideas super quick
you can do this free on scenario, under models, and it doesn't take an entire video to explain it.. its intuitive
is scenario the gaming asset platform?
make a tutorial
@@drendelous put ten images in press send
Thanks for this. QUESTION: How do you have multiple trained models in one shot? More specifically, is there a way to train several individual models, then bring them all together into one shot?
In the future you will create movies starring yourself!
This was so easy, thank you! Less steps than other tutorials.
Can it train nsfw images ?
I would like to create portraits for my family members, can I make multiple image models in the same account in regenerate? Just by uploading their images and using different names?
Amazing how AI still hasn't incorporated AI help (like Cursor, Zed etc.) in these billion parameter tools :)
great video as always Igor, commenting for the algorithm
thank you for that :)
That's fantastic, I was able to make it work without too much effort and the result are pretty good. Thanks for the great video.
This is not really playing ball with logos though... Can't manage to get this to work.
The outputs are 1MP. Any baked in or easy upscalers to take it to 4MP?
Do you know if the model is available to run it in local? Maybe in the hugging face?
Thanks for the video btw, as usual, very insightful and interesting, well done ✌️
Where can I find your AI news you can use show??
how to change references of photo of our model after it's done ?
Hi. Can I use a LoRA trained on Replicate with other accounts? In other words, can I train my LoRA with my face or a consistent character and share it for others to use with their accounts? If so, how does the process work? Thanks a lot
very cool...I'll have to try this, thanks for the video!
I got an error when training the Lora saying "Training failed. Error no file named config.json found in directory FLUX.1-dev/transformer." What do I do? HELP!
Great video! Thank you! I did the training - it completed, but I can't get it to give images. in the output it gives me an empty zip folder instead of images. I'm baffled. Although in my setup, I was never prompted for a trigger word. 🤷🏽♂
This is the worst nightmare for parents that ever said, “What do you mean-I never did that!!!” 🤣
cool but are there any recommendations for photos? how much could it cost?
hi bro ur tutorial so easy to understand, i have already make my own model, but can i use my model on my local flux dev? using comfyui?
Hi any suggestions on why Replicate will not accept my Zip File. I tried your same naming format photo_James (1) and I noticed you used PNG's when there website suggested jpg's. I tried both and both files would not upload, nor did a drag to copy. TIA, James
Thank you, your prompt template was very useful.
you didnt used hf to store the lora
Yes I wanted to keep this is simple to follow as possible and avoid technical terminology. Replicate stores the model within your account and you can use it on there or am I missing something?
@@aiadvantage nah nothing is missing, but i think we can use that lora in other models as well. well it is just a wild guess!
What base model is this based on i.e. Juggertnaut, RealisticVision? Once model has been trained, does anyone know how to use downloaded weights in another stable diffusion platform like ComfyUI?
I tested the video you recommended earlier (you had seen this video before) and it didn't work. Now I did as you taught and it worked very well, thank you!
glad to hear that it helped :)
Does this method work for training a LoRA on something like a product, car or piece of clothing?
yes absolutely. todays upload shows an example of that
Thanks! Exactly what I needed!
Hello, I can dowload the Lora for using locally? Thanks!
Is not working locally ?
is there anything similar for prompt to video
how many image can be created?
how much does it cost? can you take the model with you and generate locally or on another provider?
I put the cost into the video... cents. And no this is online as it runs on replicate gpus but yes with another workflow you could train a local model with you own images
thx for sharing. it worked :-) for me as a selfie denier a great help to create content for social media!
Help
After enter start training. It's shows logs are not available
Thanks, but I don't like the outcome personally. I wish there was a way to do it with Flux, it seems much better than Lora
wait. are they the same thing? lol
It does use Flux as the model. Lora are the images you add.
What can i do if it doesnt detect the file i upload as an imput? already tried with jpeg and png files, then created the zip file to upload, but it keepds empty
I had the same problem. @mikeys4u comment above helped me. It's the browser. Firefox didn't work for me or for him. Chrome worked for him, I'm using Edge
*That is so cool! Thanks for sharing!* 🙏 Definitely using it for some of my YT thumbnails. Danke & Beste Grüße! -Sven
Nice one man! New subscriber ✌🏻
Welcome on board :)
absolutely love this, however, after I trained myself as my nickname Halligator, most of the images comes out as an alligator in stead of the trained images of me :) Any way to change the promt name? or do I have to train a new model?
Oh no sorry to hear that but then also that's kinda funny. You will have to train a new model.
Good learning here for others: the name should be unambiguous.
The invite got expired which is linked on matt wolf's video , is there any other link for same ?
no, the free credits expired by now sorry.
This is awesome photoshoots are going to be so cheap now. Just make your own model and your set LOL
Is there concern about Replicate or Fal storing your likeness or generated images?
I heard 25 mins to prepare and would click off but since from Igor and clear walk they I’m IN!
Hay bro can we do the same thing with Fooocus Ai
why ostris?
I can't seem to get the $10 credit, does anybody know if it is still working?
No seems to be over
I enjoy your videos 💪😎 TY!
Thanks :)
Amazing with chat gpt prompt you can be whatever you want
it wont let me upload a zip file any help?
I had the same problem. @mikeys4u comment above helped me. It's the browser. Firefox didn't work for me or for him. Chrome worked for him, I'm using Edge.
Have you solved it? I also have the same problem.
Baller, thanks!
3:37 is the time stamp where i laughed out really loud😂😂- so funny✌️. Include humour it would engage more people really
can i do it for free in MacBook m1?
no more professional photography before applying for a job
Fair enough. Don't think anyone would notice
after few image it again asked payment method
$10 credit is no longer valid
Once you’ve downloaded the trained Lora can you se it in other programs locally?
10$ credit link is not working anymore
ah darn it, lasted for a month though
The upload zip of images does not seem to work in Firefox - Chrome only.
Thank you. I was using Firefox and no go but your comment has me using Edge right now and it's working.
not work for me
Done it. Its great. But it only ever generates me. So when i tell to generate a goofy putin image, it generates putin with my face. Even when i dont mention that myself... anyone with similar experiences ?
BRUUUH because u trained the ai with YOUR face
no. in the video he says it only createa images of you when you use the trigger word. in my case it always creates me. even if i don't use the trigger in my prompt.
What size is best for the upload images? Can we do 1024 or larger?
U really look like our klaus in Umbrella☂️ academy.
How can i get free 10 dollars
Free but it cost 10$
IGOR YOU SHOULD HAVE POSTED IT YESTERDAY!!! Would have saved me lots of time with my recent thumbnail 😭😭😭
Anyhow, talk to you soon! (finishing up my BSc on Sep 9th, and then my brain is finally available for other stuff! 😀)
oh no I failed you :D And congrats with the studies, looking forward to some streams together or something
results in fact quite poor with this. midjourney does that way better
I would recommend you give it another shot. This is head and shoulders above midjourneys ability to recreate humans accurately
does it provide nsfw content?
Couldn't use this news! 🧐
Why not :) Everyone can try this