Software Engineering Courses (SECourses)•3y ago

you can't copy-paste cmd lines from description.. it copies like this: pip install --no-deps --force

you can't copy-paste cmd lines from description.. it copies like this: pip install --no-deps --force-reinstall https://github.com/ArrowM/xformers/re...

ZeroCool22•3/18/23, 5:27 AM

Don't upgrade to Pytorch 2 if you have a 1080/TI GPU!

You will need to delete your VENV folder and download it again to get it generating images again.

message.txt8.17KB

!〆؂ゟトλ𝓺•3/18/23, 5:28 AM

about the new Torch2 video; if I have a GTX1060 will I still get the performance gains?

Stevens•3/18/23, 6:14 AM

Nice to meet you all. I just got into this hobby this week and an digging the videos. Really helps a ton. That said, this seems bad!

FFORTUNE VARO Hi! Thanks for the new tutorial I followed and now I´m getting this errors. Web...

Furkan Gözükara SECourses•3/18/23, 9:49 AM

hi welcome. after you changed requirements version please delete venv folder and restart. it should install all. and dont forget --xformers in command line arguments

SStevens Nice to meet you all. I just got into this hobby this week and an digging the v...

Furkan Gözükara SECourses•3/18/23, 9:49 AM

welcome

Furkan Gözükara SECourses•3/18/23, 9:49 AM

dont forget --xformers in command line arguments

Furkan Gözükara SECourses•3/18/23, 9:49 AM

restart and if still not fixed delete venv folder. it should get fixed

II A 2) Problem: launching python relauncher.py takes FIVE MINUTES! When just yesterd...

Furkan Gözükara SECourses•3/18/23, 9:51 AM

you must have a prompt error somewhere. can you show screenshot of settings

!!〆؂ゟトλ𝓺 about the new Torch2 video; if I have a GTX1060 will I still get the performance...

Furkan Gözükara SECourses•3/18/23, 9:52 AM

i think 1xxx series are not supported yet. several people reported that. i am looking for supported card list.

ZZeroCool22 ⚠ Don't upgrade to **Pytorch 2** if you have a **1080/TI GPU**! You will need t...

Furkan Gözükara SECourses•3/18/23, 9:52 AM

thanks looks like not supported yet. i am looking for supported card list. looks like starting from 2xxx cards

Nnovaaftermath it works when the yaml is removed from the directory

Furkan Gözükara SECourses•3/18/23, 9:52 AM

probably your yaml file was incorrect

KKJ you can't copy-paste cmd lines from description.. it copies like this: pip insta...

Furkan Gözükara SECourses•3/18/23, 9:53 AM

yes you need to use this page i linked that there too : https://github.com/d8ahazard/sd_dreambooth_extension/releases/tag/1.0.13

GitHub

Release 1.0.13 - Fixes and prepping for Torch 2 · d8ahazard/sd_drea...

Mar 15, 2023 Patch Notes (#1070)

Bug fixes

UI Cleanup | The txt2img checkbox has been replaced with dropdowns on the top of the Generate tab. + other tweaks.

Torch 2!
This extension is compat...

FFractal Is realistic vision 1.3 also baked? I did several tests with Dreambooth but neve...

Furkan Gözükara SECourses•3/18/23, 9:54 AM

i can say all public models are baked

Furkan Gözükara SECourses•3/18/23, 10:07 AM

haha translating my subtitles to different ones :d

Furkan Gözükara SECourses•3/18/23, 10:07 AM

first started with Spanish. do we have any one here knows Spanish?

CPI_MrRagingBull 🇬🇧•3/18/23, 10:15 AM

@Furkan Gözükara SECourses I've noticed in the new video you are python 3.10.8. I'm still running 3.10.6. Any benefit for upgrading or not really

CCPI_MrRagingBull 🇬🇧@Furkan Gözükara SECourses I've noticed in the new video you are python 3.10.8. ...

Furkan Gözükara SECourses•3/18/23, 10:19 AM

I think they are same. But I didnt have any problem with 3.10.8 yet

Furkan Gözükara SECourses•3/18/23, 10:21 AM

guys please upvote if you see this thank you : https://www.reddit.com/r/StableDiffusion/comments/11u8y8d/how_to_install_new_dreambooth_torch_2_on/

FFurkan Gözükara SECourses guys please upvote if you see this thank you : https://www.reddit.com/r/StableDi...

Viet•3/18/23, 11:42 AM

Is it work for Runpod? please !

VViet Is it work for Runpod? please !

Furkan Gözükara SECourses•3/18/23, 11:45 AM

i tried to compile torch 2 on runpod but failed. we need a unix expert to compile it :/

BillyBishkon•3/18/23, 1:06 PM

I'm getting
[!] xformers version 0.0.16rc425 installed.
Launch errors detected:
Incorrect version of xformers installed.

what version does it need? its not listed in the dreambooth requirements.txt

BBillyBishkon I'm getting [!] xformers version 0.0.16rc425 installed. Launch errors detected...

Furkan Gözükara SECourses•3/18/23, 1:09 PM

looks like they removed it from there weird. if you didnt install torch 2 add this to requirements file xformers==0.0.17.dev464

BillyBishkon•3/18/23, 1:10 PM

thank you doc

ZZono50 yeah right now i'm doing 0.005:200, 0.002:500, 0.001:3000, 0.0005 but 30% less...

CPI_MrRagingBull 🇬🇧•3/18/23, 1:57 PM

@Zono50 So after updating and installing the remove background extensions. My textual inversion training looks to be very weird. Its training really slow and im not sure if i have done something wrong. i am using your learning rate 0.005:200, 0.002:500, 0.001:3000, 0.0005.

Seb•3/18/23, 2:01 PM

i tried to follow this video https://www.youtube.com/watch?v=mnCY8uM7E50&list=PL_pbwdIyffsmclLl0O144nQRnezKlNdx3
step by step to create my own avatar. however it seems to have a few deviation from your video, not sure what went wrong.
Was using the colab file provided at comment and followed exactly every steps, the deviation as follow:
It doesnt have the section for xformer installation
It generate doggy photos from the last saved weights

Please help

YouTubeSECourses - Software Engineering Courses

Transform Your Selfie into a Stunning AI Avatar with Stable Diffusi...

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...

Furkan Gözükara SECourses•3/18/23, 2:07 PM

To downgrade back to older Torch first delete venv, let it reinstall, then activate venv and run this command pip install -r "path_of_SD_Extension\requirements.txt"

SSeb i tried to follow this video https://www.youtube.com/watch?v=mnCY8uM7E50&list=PL...

Furkan Gözükara SECourses•3/18/23, 2:07 PM

1 : xformers is included in another script now. dont worry about that

Furkan Gözükara SECourses•3/18/23, 2:08 PM

2: doggy photos meaning that your prompt is not correct. show settings i can tell you

FFurkan Gözükara SECourses To downgrade back to older Torch first delete venv, let it reinstall, then activ...

CPI_MrRagingBull 🇬🇧•3/18/23, 2:18 PM

Sorry. You recommending this to my issue

FFurkan Gözükara SECourses 2: doggy photos meaning that your prompt is not correct. show settings i can tel...

Seb•3/18/23, 2:22 PM

Thank you for your quick reply!

This is the setting I have, followed almost exactly from your video:
!accelerate launch train_dreambooth.py \
--pretrained_model_name_or_path=$MODEL_NAME \
--pretrained_vae_name_or_path="stabilityai/sd-vae-ft-mse" \
--output_dir=$OUTPUT_DIR \
--revision="fp16" \
--with_prior_preservation --prior_loss_weight=1.0 \
--seed=1337 \
--resolution=512 \
--train_batch_size=1 \
--train_text_encoder \
--mixed_precision="fp16" \
--use_8bit_adam \
--gradient_accumulation_steps=1 \
--learning_rate=1e-6 \
--lr_scheduler="constant" \
--lr_warmup_steps=256 \
--num_class_images=384 \
--sample_batch_size=4 \
--max_train_steps=2560 \
--save_interval=10000 \
--save_sample_prompt="photo of zwx dog" \
--concepts_list="concepts_list.json"

Seb•3/18/23, 2:23 PM

there are some errors that doesnt seems to appears on your video too:
The following values were not passed to accelerate launchaccelerate launch and had
To avoid this warning pass in values for each of the problematic parameters or run accelerate configaccelerate config.
2023-03-18 13:50:17.755195: I tensorflow/core/platform/cpu_feature_guard.cc:193] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: AVX2 FMA
To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.
2023-03-18 13:50:20.351469: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer.so.7'; dlerror: libnvinfer.so.7: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/lib64-nvidia
2023-03-18 13:50:20.351702: W tensorflow/compiler/xla/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libnvinfer_plugin.so.7'; dlerror: libnvinfer_plugin.so.7: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /usr/lib64-nvidia
2023-03-18 13:50:20.351727: W tensorflow/compiler/tf2tensorrt/utils/py_utils.cc:38] TF-TRT Warning: Cannot dlopen some TensorRT libraries. If you would like to use Nvidia GPU with TensorRT, please make sure the missing libraries mentioned above are installed properly.
/usr/local/lib/python3.9/dist-packages/accelerate/accelerator.py:243: FutureWarning: logging_dirlogging_dir is deprecated and will be removed in version 0.18.0 of

Accelerate. Use project_dirproject_dir instead.
warnings.warn(

Seb•3/18/23, 2:23 PM

/usr/local/lib/python3.9/dist-packages/bitsandbytes/cuda_setup/paths.py:105: UserWarning: /usr/lib64-nvidia did not contain libcudart.so as expected! Searching further paths...
warn(
/usr/local/lib/python3.9/dist-packages/bitsandbytes/cuda_setup/paths.py:27: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('/sys/fs/cgroup/memory.events /var/colab/cgroup/jupyter-children/memory.events')}
warn(
/usr/local/lib/python3.9/dist-packages/bitsandbytes/cuda_setup/paths.py:27: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('//colab.research.google.com/tun/m/cc48301118ce562b961b3c22d803539adc1e0c19/gpu-t4-s-2rmy4vqcs4arz --tunnel_background_save_delay=10s --tunnel_periodic_background_save_frequency=30m0s --enable_output_coalescing=true --output_coalescing_required=true'), PosixPath('--listen_host=172.28.0.12 --target_host=172.28.0.12 --tunnel_background_save_url=https')}
warn(
/usr/local/lib/python3.9/dist-packages/bitsandbytes/cuda_setup/paths.py:27: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('/env/python')}
warn(
/usr/local/lib/python3.9/dist-packages/bitsandbytes/cuda_setup/paths.py:27: UserWarning: WARNING: The following directories listed in your path were found to be non-existent: {PosixPath('//ipykernel.pylab.backend_inline'), PosixPath('module')}
warn(
CUDA_SETUP: WARNING! libcudart.so not found in any environmental path. Searching /usr/local/cuda/lib64...

CCPI_MrRagingBull 🇬🇧Sorry. You recommending this to my issue

Furkan Gözükara SECourses•3/18/23, 2:24 PM

if your graphic card is now 2000+ yes

SSeb Thank you for your quick reply! This is the setting I have, followed almost exa...

Furkan Gözükara SECourses•3/18/23, 2:24 PM

,your save sample is

Furkan Gözükara SECourses•3/18/23, 2:24 PM

--save_sample_prompt="photo of zwx dog" \

Furkan Gözükara SECourses•3/18/23, 2:24 PM

you see

FFurkan Gözükara SECourses if your graphic card is now 2000+ yes

CPI_MrRagingBull 🇬🇧•3/18/23, 2:24 PM

Im using RTX 3060 12GB with Auto1111

Furkan Gözükara SECourses•3/18/23, 2:24 PM

also i have more up to date video that will help you tremendously : https://youtu.be/kIyqAdd_i10

YouTubeSECourses - Software Engineering Courses

Stable Diffusion Google Colab, Continue, Directory, Transfer, Clone...

Our Discord : https://discord.gg/HbqgGaZVmr. This is the video where you will learn how to use Google Colab for Stable Diffusion. 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...

CCPI_MrRagingBull 🇬🇧Im using RTX 3060 12GB with Auto1111

Furkan Gözükara SECourses•3/18/23, 2:24 PM

use torch 2 better

FFurkan Gözükara SECourses use torch 2 better

CPI_MrRagingBull 🇬🇧•3/18/23, 2:25 PM

So that is what i am using but should my TI be learning at the slow rate?

FFurkan Gözükara SECourses also i have more up to date video that will help you tremendously : https://yout...

Seb•3/18/23, 2:26 PM

ok thanks, will give it a try again!

AlaBeat•3/18/23, 2:28 PM

hi when i try to update to torch 2 i get this ERROR: Could not install packages due to an OSError: [WinError 5] Access is denied: 'C:\Users\me\stable-diffusion-webui\venv\Lib\site-packages\~orch\lib\asmjit.dll'
Check the permissions.

CCPI_MrRagingBull 🇬🇧So that is what i am using but should my TI be learning at the slow rate?

Furkan Gözükara SECourses•3/18/23, 2:30 PM

so your TI slower with torch 2? interesting maybe it is because automatic1111 still didnt move torch 2 yet

AAlaBeat hi when i try to update to torch 2 i get this ERROR: Could not install packages ...

Furkan Gözükara SECourses•3/18/23, 2:31 PM

run cmd as administrator

Furkan Gözükara SECourses•3/18/23, 2:31 PM

also i would install web ui directly into C drive not inside users folder

AlaBeat•3/18/23, 2:32 PM

problem is i already running out of space and will delete sd folder at some point and move it to another directory, but until my new ssd arrives i will still keep it there

FFurkan Gözükara SECourses so your TI slower with torch 2? interesting maybe it is because automatic1111 st...

CPI_MrRagingBull 🇬🇧•3/18/23, 2:33 PM

Yep. after watching your new video I updated to the new torch 2.0. I am using the learning rate of 0.005:200, 0.002:500, 0.001:3000, 0.0005.

CCPI_MrRagingBull 🇬🇧Yep. after watching your new video I updated to the new torch 2.0. I am using th...

Furkan Gözükara SECourses•3/18/23, 2:33 PM

i didnt try such learning rate. the constant one was good for me

FFurkan Gözükara SECourses i didnt try such learning rate. the constant one was good for me

CPI_MrRagingBull 🇬🇧•3/18/23, 2:34 PM

I will try with the default 0.005 learning rate and see if that makes a difference

FFurkan Gözükara SECourses i didnt try such learning rate. the constant one was good for me

CPI_MrRagingBull 🇬🇧•3/18/23, 2:43 PM

Ok it was the learning rate. 0.005 seems to have done the trick

you can't copy-paste cmd lines from description.. it copies like this: pip install --no-deps --force

Similar Threads