Hello, I have 300 steps per image, but with 10 images I have 16800 steps. What have I done wrong?
Hello, I have 300 steps per image, but with 10 images I have 16800 steps. What have I done wrong?
J
Jis there any way to solve this, this is a new error
J
Mwill a strong cpu help with ai in general? like sorting dataset and other stuff? I need to know if i should buy a 7950x3d
Mtarget is to also train an ai for playing games
FYes
FGitHub
…s started
fix #7714
Traceback (most recent call last):
File "launch.py", line 361, in <module>
start()
File "launch.py", line 356, in start
webui.webui()
Fi...
fix #7714
Traceback (most recent call last):
File "launch.py", line 361, in <module>
start()
File "launch.py", line 356, in start
webui.webui()
Fi...
F300 x 10 x 2 for dreambooth. Make sure that you don't have sub folders inside your training filer
FFolder I am sleeping
Jwhat i need to do after download fastapi file ?
Nhi there! i already made some nice trained images via google colab and hf. then I decided to not only use automatic1111 locally for image generation but upgrade my graiccard from a 1660 to a 3060 with 12gb to be able to run dreambooth as well on my own hardware.
i was able to run some dreambboth creation but obviously there had been som error in setup.
then I tried to follow the git bash d8f8bcb821fa62e943eb95ee05b8a949317326fe command to bring my setup to the level from the last tut- and somehow killed the whole auomatic1111 instal
i was able to run some dreambboth creation but obviously there had been som error in setup.
then I tried to follow the git bash d8f8bcb821fa62e943eb95ee05b8a949317326fe command to bring my setup to the level from the last tut- and somehow killed the whole auomatic1111 instal
J
Ji tried to install fastapi but still error
Ji solved it by upgrading fastapi on python in venv folder i was not able to pass because i execute above python command in appdata folder
ADreamboth Tutorial Video Question: Hi, I followed your video tutorial but can’t get the Dreambooth tab to show up in my webui. The log shows: ModuleNotFoundError: No module named 'tensorflow.python' I have tried deleting and reinstalling the extension as well as updating to the most recent version of python. I'm also trying to start over by installing the webui to a new location on my computer but it always has errors how when I do this now. Any idea how I can best proceed? (I don’t understand any coding languages) Thanks.
CarXiv.org
State-of-the-art computer vision systems are trained to predict a fixed set
of predetermined object categories. This restricted form of supervision limits
their generality and usability since...
of predetermined object categories. This restricted form of supervision limits
their generality and usability since...
CIf I understand this correctly:
2.1 is based on LAION-5B dataset, and 1.5 is based on CLIP
1.5 is based on LAION-5B dataset containing Adult Material
Stable Diffusion v1 was trained on subsets of LAION-2B(en) which is based on primarily english descriptions
The Stable-Diffusion-v1-5 checkpoint was initialized with the weights of the Stable-Diffusion-v1-2 checkpoint and subsequently fine-tuned on 595k steps at resolution 512x512 on "laion-aesthetics v2 5+" and 10% dropping of the text-conditioning to improve classifier-free guidance sampling.
Stable Diffusion 2.x uses a machine learning model called MiDaS that’s trained on a combination of 2D and 3D image data — in particular, it was trained using a 3D movies dataset containing pairs of stereoscopic images. With this training, the model can look at a 2D scene and then back into a 3D representation of it.
Stable Diffusion 1.0 used OpenAI’s CLIP model for interpreting text prompts, but 2.x switches over to a newer model called OpenCLIP that brings a number of improvements, but also is the source of some complaints.
Both the OpenCLIP model and its predecessor are open-source, but only OpenCLIP was trained on an open database of images. This open database has a few important differences from the proprietary database OpenAI used to train the CLIP model used in SD 1.5:
It was explicitly and aggressively filtered for NSFW images
It contains fewer celebrities and named artists
Some terms that dominated CLIP to an unreasonable degree (“Greg Rutkowski,” “trending on artstation”) are now either absent or their dominance is greatly reduced.
2.1 is based on LAION-5B dataset, and 1.5 is based on CLIP
1.5 is based on LAION-5B dataset containing Adult Material
Stable Diffusion v1 was trained on subsets of LAION-2B(en) which is based on primarily english descriptions
The Stable-Diffusion-v1-5 checkpoint was initialized with the weights of the Stable-Diffusion-v1-2 checkpoint and subsequently fine-tuned on 595k steps at resolution 512x512 on "laion-aesthetics v2 5+" and 10% dropping of the text-conditioning to improve classifier-free guidance sampling.
Stable Diffusion 2.x uses a machine learning model called MiDaS that’s trained on a combination of 2D and 3D image data — in particular, it was trained using a 3D movies dataset containing pairs of stereoscopic images. With this training, the model can look at a 2D scene and then back into a 3D representation of it.
Stable Diffusion 1.0 used OpenAI’s CLIP model for interpreting text prompts, but 2.x switches over to a newer model called OpenCLIP that brings a number of improvements, but also is the source of some complaints.
Both the OpenCLIP model and its predecessor are open-source, but only OpenCLIP was trained on an open database of images. This open database has a few important differences from the proprietary database OpenAI used to train the CLIP model used in SD 1.5:
It was explicitly and aggressively filtered for NSFW images
It contains fewer celebrities and named artists
Some terms that dominated CLIP to an unreasonable degree (“Greg Rutkowski,” “trending on artstation”) are now either absent or their dominance is greatly reduced.
GUse “pip install tensorflow”
A@Gant Thanks for your response, I did that earlier but wasn't sure if I did it right, I just rebooted now and the tab is finally there, yippy!!! 
CYouTubeDaniel Bourke
Welcome to the most beginner-friendly place on the internet to learn PyTorch for deep learning.
All code on GitHub - https://dbourke.link/pt-github
Ask a question - https://dbourke.link/pt-github-discussions
Read the course materials online - https://learnpytorch.io
Sign up for the full course on Zero to Mastery (20+ hours more video) - https:/...
All code on GitHub - https://dbourke.link/pt-github
Ask a question - https://dbourke.link/pt-github-discussions
Read the course materials online - https://learnpytorch.io
Sign up for the full course on Zero to Mastery (20+ hours more video) - https:/...
Fhi welcome. i can connect your computer and do quick install
Fbut you can watch this video to do fresh install if you want
Fah i am glad you did
Fi think i did read but not remembering much atm
Fwhich image is best you think for thumbnail
F
P@Dr. Furkan Gözükara i really appreciate what you do here. i mean... you go way above normal helping, and get to the core of the problem... with everyone! i am so happy i found this discord. you've solved so many issues i've had with SD. my coworkers are all going to MJ, but i feel like it's a limitation by it's easy and simplistic output. SD is where it's going to be.
Ppleaase make larger.
Pwshatever this is
P
Phas good color balance, and nice detail. realistic background
乂2nd and 17th
Pcolor balance is super on this one.
PThis one, thouh
Pppl might be like, I WANT TO MAKE THAT
Pthe contrast of green on bottom to red on top, make ppl enjoy an image.
PYour thumb will be go from sketch to art, will you show the line art?
Fi am uploading files so you can check better
Fthis is i am explaining in this video
Pkk
乂@Dr. Furkan Gözükara sir can you please share your configuration of SD as you have the same memory GPU?
F
Fahmed can you instally any desk i can make a fresh intallation
Fthat is best wy
Pwhen i said earlier that nothing good happens in that hour.... that hours is now, here. super tired. 2 am. ( i know yours was 3:14 or something, but same,... no good decisions)
F1:02 pm here :d
Fnow
Punfair.
Fpost processing video
