Комментарии:
Hi, Thanks for your work but the github page has changed and the command lines are not there anymore, so this tutorial is useless without an update. could you do it?
ОтветитьI tried your advice to use pictures in different settings, clothing and poses. Didn't work at all. Hair cut, beard and face are always completely different.
ОтветитьDark mode please!!! my eyes!
ОтветитьMac users?
Ответитьthe image is 512 x 512 pixels or inches to be cropped?
ОтветитьThanks for all the knowledge. I’ve learned a lot from you, Olivio. It’s appreciated. Curious if LoRas can be trained with a landmark like a fountain, building or park? Cheers.
Ответитьjust a suggestion please show various results properly ( more renders, different prompts etc.) thx.
ОтветитьNot you but the procedure is too difficult and too confusing.
ОтветитьFire flame spitter. Thank you for the info!
Ответитьlora saved model as .json file instead of safestensors and i get CalledProcessError: Command '['C:\\kohya\\kohya_ss\\venv\\Scripts\\python.exe' can someone help?
ОтветитьSo i see the warning "The script uses different library versions and WILL break your webui." from others, I have no idea how to avoid this by "creating a new environment", so is the way you show safe for me to run so i do not break my automatic1111 on the same computer/hard drive?
Ответитьit's asking me if I want to use fp16, bf16 or fp8. What does that mean??
Ответитьafter last lora's update can't install lora at all. it drops a lot of errors about unfound modules 'rich, rip, torch' etc and no explanation where to get them.
ОтветитьThis tutorial is so far off its funny,
I tried correcting the instructions, but thanks to censorship, I cannot.
Has anyone else had the gui-user.bat file not appear? I do the installation but the file does not exist.
ОтветитьHow can I add trigger word for the lora?
Ответитьbro so much has changed
we need neeed a new video
It's a suit, not a suite. A suite is a set of rooms.
Ответитьplease never say these installations are "easy" it's condescending and infuriating
ОтветитьOut dated info here. The instructions are different in the link. The installation steps are different and the installation folder is also different. I cannot get this work at all! Too much mixed instructions. :(
ОтветитьHi! thank you so much for this. Question: The github link is looking different now with different steps, I think I managed but after isntalling cudann, I can't fint the upgrade.ps1. Would Be ever so grateful if anyone knows how to fix it
Ответитьthere is a sample folder created in my model folder, am i doing it right?
ОтветитьSadly this doesn't work for GPUs of lower RAM like 6GB. Anyone have any workarounds?
Ответитьwhy the blip captioning not working for me?, its showing error after loading the blip captioning without it dowloading
ОтветитьI have a RTX 4090 and with batch size 1 it took 1 hour for 4400 steps. Is there something wrong in my configuration?
ОтветитьHello Olivio, I'm a little confused, Wy need to install Additional extension for it ?
ОтветитьI followed all the steps from the video, but it doesn't use my face in a portrait. Just generates a random guy...
ОтветитьI'm hitting a wall every time I try to install Lora, All copy and paste commands never work for me.
ОтветитьThe github page looks completely different!
ОтветитьUnfortunate people make videos like these for Windows only. Quite a lot of the webui users are on Linux.
ОтветитьI was able to install everything but my lora has no effect, with or without <mylora:1>, the result is the same. What could be causing this? I used the file on the output folder
ОтветитьEverything looks different, this tutorial is no longer useful.
Ответитьme got error whhile installing the it says ERROR: Could not find a version that satisfies the requirement torch==1.12.1+cu116 (from versions: 2.0.0, 2.0.1)
ERROR: No matching distribution found for torch==1.12.1+cu116
Did Dreambooth built into Automatic1111 make this obsolete?
ОтветитьNot working on a GTX 1650. I know it's a weak card but, i would to try before to change my computer. CUDA cannot stop crashing. Maybe there is some params i can change in order to do so ? Even if the resolution is way lower, i don't care, i am just impatient to try...
ОтветитьWith BLIP Captioning i have the error FileNotFoundError: [WinError 2] idk what to do i have the pictrues i the folder but it wont work
ОтветитьI have 4Gb of vram only. Can I teain Lora on CPU?
ОтветитьI try everything, even do a 1 hour training by giving 10 epochs... nothing.
ОтветитьWhy is 9MB the file? i try whit 100 images and still 9MB
ОтветитьHello, Can I ask if there is a way to create Lora's for low end PC using the runpod services?
ОтветитьYou didnt explain how the additional network is even used. I did everything as in the video, but my model is called last_safetensors, and I have no idea why last, I didnt name it that. Anyways, I put it into the lora and all, but it has no effect whatsoever
ОтветитьDo your instructions work on linux or just DOS-based operating systems?
ОтветитьBars...
Ответитьplz new version of the video
Ответить"not the best acronym" 💀
Ответитьthose bars in the beginning were off the chain
ОтветитьHi, I have a question, how to use Lora model in Deforum ? to produce animation video ? ;) if you have a link to a tutorial also be great. I appreciate it.
ОтветитьI checked out to first install Visual Studio and get the option to choose about 5476 options there....😟
ОтветитьSmooth!😎
Ответить