TF-TRTWarning:CouldnotfindTensorRT
Original message was deleted
PHello, I tried to follow https://youtu.be/c_S2kFAefTQ and https://youtu.be/3uzCNrQao3o videos last night to install SD and Kohya SS on RunPod but ran into the following error when trying to Run BLIP captioning on a set of images. What might have been the issue?
YouTubeSECourses
I have shown how to install latest version of Automatic1111 Web UI for Stable Diffusion and DreamBooth extension of Auto1111 on RunPod in this video. Moreover, I show how to upgrade to latest Cuda, Torch and cuDNN DLL files. With these upgrades the image generation speed literally doubles.
GitHub Readme File
https://github.com/FurkanGozukara...
GitHub Readme File
https://github.com/FurkanGozukara...
YouTubeSECourses
How to install famous Kohya SS LoRA GUI on RunPod IO pods and do training on cloud seamlessly as in your PC. Then use Automatic1111 Web UI to generate images with your trained LoRA files. Everything is explained step by step and amazing resource GitHub file is provided with necessary commands. If you want to use Kohya's Stable Diffusion trainers...
FLooks like worked what is the problem
FAh it couldn't find any images
FI would try to pre process images with stable diffusion web ui training tab
FI have shown in last video
FIt says found 0 images
PYea, it says 0 images found in a folder of 20 .jpeg images. I even tried creating another folder with 20 .jpg images in case I had the wrong file type, but I got the same errors running several of the auto captioning tools via the Kohya SS gui.
PUnder the dreambooth training tab? Your lastest dreambooth Photo studio video?
Fit is under training tab of automatic1111 web ui itself
PWhen I copied path via Jupiter labs it pasted as “workspace/HalBot” instead of “/workspace/HalBot” …. I was missing a “/“ at the beginning
Fyep i show this in every video. they still didnt fix this mistake
PYes, thank you. I was rewatching part of one of your video and finally caught it.
But now, I have another error after the start of training. It looks to be regarding the folder names.
Prior to this step, I’ve found success with the auto captioning, and creation of 900 regularization images. My training images are 1280 max on longest side, I may drop that down to 768 or 512 until I get my bearings with the process.
But now, I have another error after the start of training. It looks to be regarding the folder names.
Prior to this step, I’ve found success with the auto captioning, and creation of 900 regularization images. My training images are 1280 max on longest side, I may drop that down to 768 or 512 until I get my bearings with the process.
P
P@Dr. Furkan Gözükara
PI am following this to the best of my ability, but something isn't working.
https://github.com/bmaltais/kohya_ss/blob/master/docs/image_folder_structure.md
https://github.com/bmaltais/kohya_ss/blob/master/docs/image_folder_structure.md
GitHub
Contribute to bmaltais/kohya_ss development by creating an account on GitHub.
PI gave it a solid troubleshoot session last night before calling it a night with my progress.
Any pointers or help is appreciated but not expected. Thank you for sharing as much as you have. I don't think I can even describe myself competent at html and css, so again, thank you, I couldn't have gotten this far without your guidance via YT tutorials!
Any pointers or help is appreciated but not expected. Thank you for sharing as much as you have. I don't think I can even describe myself competent at html and css, so again, thank you, I couldn't have gotten this far without your guidance via YT tutorials!
P@Dr. Furkan Gözükara
Ptkinter.TclError: no display name and no $DISPLAY environment variable
PI'm curious if this is the culprit?^
PGitHub
Traceback (most recent call last): File "/data/7082/venv/lib/python3.8/site-packages/gradio/routes.py", line 399, in run_predict output = await app.get_blocks().process_api( File "/d...
PTry starting the gui with --headless as an option... for example:
.\gui.sh --headless
.\gui.sh --headless
Flet me test my installer.
Fjust installed with my script
Fkohya started
Fgonna test now
FGitHub
Best Stable Diffusion and AI Tutorials, Guides, News, Tips and Tricks - Stable-Diffusion/Tutorials/How-To-Install-Kohya-LoRA-Web-UI-On-RunPod.md at main · FurkanGozukara/Stable-Diffusion
Fjust tested working very well
PGreat to know. My first attempt with the “/“ error was following your manual install instructions. It took me nearly 2.5 hours to complete the installation process. I was very frustrated by the “/“ error and found some info stating that I should just build from scratch, so I deleted the pod.
I’ve now set up a pod with Ashleykza template, and that’s where I’m experiencing the current error above.
I’ve now set up a pod with Ashleykza template, and that’s where I’m experiencing the current error above.
Fi dont know that error unfortunately but my manual installation steps are also updated
Fyou can also download scripts from here https://www.patreon.com/posts/84898806
PGreat, I’ll take a look at the new manual install steps and try to understand what’s going on / what has changed. I’m using chatGPT to interpret all the commands to better understand what the heck I’m doing.
Again, thank you so much for putting out the videos. They have helped me out a great deal.
If you’d consider it, as a beginner, I would immensely appreciate it if you could briefly describe what each command is doing. Some commands are straightforward like downloading, but some are more complex.
Also, as a beginner, I found it difficult to follow the manual installation process while you switched back and forth from the automatic process. The time stamps helped, but it would be most helpful to see an uncut and linear installation process for both (with speed alterations that fast forward through installation waiting periods). For example, “here is the manual installation, but if all that is too much, skip to 6:00 to view my how easy it is to use my automatic script to save time.”
Again, thank you so much for putting out the videos. They have helped me out a great deal.
If you’d consider it, as a beginner, I would immensely appreciate it if you could briefly describe what each command is doing. Some commands are straightforward like downloading, but some are more complex.
Also, as a beginner, I found it difficult to follow the manual installation process while you switched back and forth from the automatic process. The time stamps helped, but it would be most helpful to see an uncut and linear installation process for both (with speed alterations that fast forward through installation waiting periods). For example, “here is the manual installation, but if all that is too much, skip to 6:00 to view my how easy it is to use my automatic script to save time.”
Fi can give special consultation if you are interested in such thing. 1 on 1 explanation
PThank you. It turned out that it wanted my image folder in the kohya SS folder. Maybe a complication with the venv? I also had an error linking to a copy of SD v1-5 in any location. I had to use the quick selection for the training model. BUT, I was able to successfully train my first Lora with great results! 