are you doing this on kohya ss? what is your repeating count?
are you doing this on kohya ss? what is your repeating count?
Fi prefer to not use it if possible because i remember there once was a bug with it which prevented training
Prpg_v4 + controlnet + multi_frame video rendering
Pfrom your tutorial https://youtu.be/kmT-z2lqEPQ
YouTubeSECourses
RunPod: https://bit.ly/RunPodIO. Discord : https://bit.ly/SECoursesDiscord. Turn your real video into an animation with just 1 click by using Stable Diffusion for free. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on 
https://www.patreon.com/SECourses
Playlist of Sta...
https://www.patreon.com/SECoursesPlaylist of Sta...
Pusing a custom LoRa and going for more photoreal
Pworking on a new render now to add more facial expression using two controlnets
RSo good
Pthank you. it's been a loooong day.
Pit'll be 6 hours till the render is done for the new video. fingers crossed for a touch of temporal coherence
THey had the same yesterday after watching the tutorial. Have you looked at the prompt of the generated images?
I saw only by chance later that there were words in my prompt that I did not enter
This was due to the extension Unprompted, which I once installed. This added keywords like nude and modelshoot
I saw only by chance later that there were words in my prompt that I did not enter
This was due to the extension Unprompted, which I once installed. This added keywords like nude and modelshoot
TEven after deactivation the prompt was added, so I had to restart automatic
Rwhat kind of progress are you making with temporal coherence, is this a holy grail, or do you see smoothness in the horizon? Just wondering if its readily available provided time is invested. But what is the timeframe, a couple of days, weeks? mad scientist dedication?
RJust wondering if worth getting into for myself
Fautomatic1111 released a new release that is good
PI'm getting okay results. Really hit and miss depending on how complex the generation is. I'm by no means an expert at this stuff so it's a lot of trial and error. As for when we'll achieve true temporal coherence? Could be tomorrow could be next year. But I'm sure there's people much smarter than me working on it right now. I'm trying to learn it because I'm setting goals for myself with stable diffusion. Rather than just endlessly generating images. I want to progress with the tech.
MI keep getting errors like,
CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 10.76 GiB total capacity; 4.29 GiB already allocated; 10.12 MiB free; 4.46 GiB reserved in total by PyTorch)
I have a 8gb 3070ti, is there any reason for me not to be able to use Konya correctly?
CUDA out of memory. Tried to allocate 20.00 MiB (GPU 0; 10.76 GiB total capacity; 4.29 GiB already allocated; 10.12 MiB free; 4.46 GiB reserved in total by PyTorch)
I have a 8gb 3070ti, is there any reason for me not to be able to use Konya correctly?
Rgood to know. it's good to focus on things, to not get caught up in the noise of everything that is happening
PFor sure
PIs this the fabled update that supports torch2.0?
Fyep
Fit was already supporting
Fi already have shown multiple times
Fbut now out of box
PYeah. I followed one of your tutorials that had torch2.0. I got it set up and then a git pull reverted me right back. Lol. Didn't have the heart to go through that again. I'll update in the morning when my render is done.
Fah i see what you mean. i dont do automatic updates D:
Fwould you guys be interested in deep voice cloning?
TI'm pulling my hair out here, where does the calculation come from that you see in the text file video????
https://youtu.be/mnCY8uM7E50?t=1153
https://youtu.be/mnCY8uM7E50?t=1153
YouTubeSECourses
Our Discord : https://discord.gg/HbqgGaZVmr. How to do free Stable Diffusion DreamBooth training on Google Colab for free. If I have been of assistance to you and you would like to show your support for my work, please consider becoming a patron on https://www.patreon.com/SECourses
Playlist of Stable Diffusion Tutorials, Automatic1111 and Goo...
https://www.patreon.com/SECoursesPlaylist of Stable Diffusion Tutorials, Automatic1111 and Goo...
Fit is totally based on experience of users
Fbut the best way is
TI tried to get ChatGPT to explain dreambooth_training.py to me.
So far this worked well.
Then I asked what values I should put here, starting from 15 images of mine:
lr_warmup_steps, num_class_images, sample_batch_size, max_train_steps,save_interval
After analyzing all the code, I thought she should know on what bases the calculation is based. But apparently it doesn't
According to the calculation in the video it should be 15*12=180. But ChatGPT says 15...
Furthermore she says:
The number of class images --num_class_images should be set to 12, because this is the number of training images you have for your class.
The --sample_batch_size depends on how much GPU memory you have available. A batch size of 4 is relatively small and should be fine for most GPUs, but it depends on your specific configuration.
The --max_train_steps should be at least large enough to go through all the training frames once. Since you have 12 training frames and use a batch size of 1, you should set at least 12 steps, e.g. --max_train_steps=12.
The --save_interval depends on how often you want to save the model weights. A good rule is to save every few epochs. Since you only have 12 training frames, you might not want to train too many epochs to avoid overfitting. For example, if you want to train 3 epochs, you can use --max_train_steps=36 and --save_interval=12 to save the model after each epoch.
So far this worked well.
Then I asked what values I should put here, starting from 15 images of mine:
lr_warmup_steps, num_class_images, sample_batch_size, max_train_steps,save_interval
After analyzing all the code, I thought she should know on what bases the calculation is based. But apparently it doesn't
According to the calculation in the video it should be 15*12=180. But ChatGPT says 15...
Furthermore she says:
The number of class images --num_class_images should be set to 12, because this is the number of training images you have for your class.
The --sample_batch_size depends on how much GPU memory you have available. A batch size of 4 is relatively small and should be fine for most GPUs, but it depends on your specific configuration.
The --max_train_steps should be at least large enough to go through all the training frames once. Since you have 12 training frames and use a batch size of 1, you should set at least 12 steps, e.g. --max_train_steps=12.
The --save_interval depends on how often you want to save the model weights. A good rule is to save every few epochs. Since you only have 12 training frames, you might not want to train too many epochs to avoid overfitting. For example, if you want to train 3 epochs, you can use --max_train_steps=36 and --save_interval=12 to save the model after each epoch.
Fhaving checkpoints and comparing them
Fwhat gpu do you have?
TTrying the Google Colab currently.
TOn my local machines a RTX2080 Ti and on my Laptop RTX3070 8GB. I searching for a good eGPU
TOn Colab its the T4 - I used the link from the video
P3070 8gb is pretty good for local training. Just make sure you have a good cooling pad for the laptop.
TI can use Dreambooth on the RTX2080 - I tested it, but my results all where bad
Not because of the machine but because of my knowledge
Not because of the machine but because of my knowledge
PWhat are you trying to train? Yourself?
TYes at first myself and my son. And afterwards some Web UI
Fwell you have much better options
Fwith rtx 2080 because you can easily save checkpoints
Fand compare them
Fhave you installed dreambooth?
Fi can quickly install dreambooth on your machine
PYouTubeSECourses
#Kohya SS web GUI DreamBooth #LoRA training full tutorial. You don't need technical knowledge to follow this tutorial. In this tutorial I have explained how to generate professional photo studio quality portrait / self images for free with Stable Diffusion training.
Our Discord server
https://bit.ly/SECoursesDiscord
If I have been of assis...
Our Discord server
https://bit.ly/SECoursesDiscord
If I have been of assis...
TYes I made that. But all of them are pretty bad. So I was wondering if it is because of my settings in the Colab.
Fye these settings also working great
PFor sure
F256 network size works even better
Fare you interested in learning fundementals?
TThe Kohya SS is running currently under my RTX2080. Im looking the video on my TV
But besides I wanted to start a training in the colab. So I can compare the LoRA and things I made with "standard" Dreambooth..
But besides I wanted to start a training in the colab. So I can compare the LoRA and things I made with "standard" Dreambooth..