Hello everyone. I am Dr. Furkan Gözükara. PhD Computer Engineer. SECourses is a dedicated YouTube channel for the following topics : Tech, AI, News, Science, Robotics, Singularity, ComfyUI, SwarmUI, ML, Artificial Intelligence, Humanoid Robots, Wan 2.2, FLUX, Krea, Qwen Image, VLMs, Stable Diffusion
Yes, in your tutorial it was clear. On Runpod I can't tell which one is supported, so I always run the wizard. Yesterday the wizard was always wrong, for some reason. On my local machine, the wizard gives me nearly the best settings! So my only solution was to train, cancel, see if the images were black or not. It took me hours to find the right settings, but it's time consuming. Then on a different pod, it's the same story again, which means paying for extra hours.
(Runpod experience) Speaking of extra hours, runpodctl download speed is...honestly very disappointing. 3-5 MB/s in the best case, seriously? And it interrupts very often. My line can handle 100 MB/s, I never had one interruption in a year of work. I use it everyday, so I know for sure that the issue is not from my end, as I use plenty of different sources to do transfer. For instance, installing Automatic 1111 from zero, with all the packages and models is a thing I can do in less than 1 hour. This has nothing to do with Mr. Furkan's tutorials, but since he suggested to use runpodctl, I am sharing my experience. It could be an isolted case, I m not sure. So I begin to think that runpod just want us to sped more hours, if you add the fact that gradio freezes every single time, and the fact that often times GPUs are assigned when you are idle, which means again > transferring files > extra hours.
In short, using runpodctl to move 1 GB TO Runpod takes 10 second, which is fantastic. Moving it back to the same source takes 15 minutes if you re lucky. Moving the same size from another source (not from runpod) to the same destination takes less time, which proves that the transfer speed is controlled by Runpod.
Great service for GPU, but given that I am not an expert someone please explain why runpod's download speed is so terrible. I get faster download via browser,but...they interrupt.
PyTorch 2.0.0 is now GA in the last 24 hours and has the cuDNN v8.7 fix if you get the correct version of it. In other words, no more file copying hacks. However, there are two versions of 2.0.0. t...
anyone have any tips on choosing a pod with enough vram? I keep choosing ones that have more than 48 for instance but then it shows it is only running at 17? does that make sense? my training samples come out black boxes. I just had a great pod the past 5 or 6 days running well but it was scheduled for maintenance and forgot my selection.
Finally I was able to training my face and get a checkpoint which is like me. But any prompt I putt on the text box in tx2img dont change my image at all.
Hi! Thanks for the new tutorial I followed and now I´m getting this errors. WebUI opens and works but Dreambooth doesn´t appear ####################################################################################################### Initializing Dreambooth If submitting an issue on github, please provide the below text for debugging purposes:
Python revision: 3.10.6 (tags/v3.10.6:9c7b4bd, Aug 1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)] Dreambooth revision: 8e7277b7ec30801753e3591e0e1139b97ec88e56 SD-WebUI revision:
Checking Dreambooth requirements... Failed to install Dreambooth requirements. [+] torch version 1.13.1+cu117 installed. [+] torchvision version 0.14.1+cu117 installed. [!] xformers version 0.0.16rc425 installed. [+] accelerate version 0.17.1 installed. [+] bitsandbytes version 0.35.4 installed. [!] diffusers NOT installed. [+] transformers version 4.27.1 installed. Launch errors detected: Incorrect version of xformers installed. diffusers not installed.
1) It is using less VRAM than 2 days ago Until 2 days ago, I was only able to work with 48 VRAM. Now I'm using 24 GB, with heavy settings, and it uses 15-18 GB. With light settings even less than 6 GB.
2) Problem: launching python relauncher.py takes FIVE MINUTES! When just yesterday it was 30 seconds
Consideing that Gradio freezes every single time, this is increasing time usage for users on Runpod. I already sort of complained about slow download speed on Runpod's channel, and now we have this to make things even slower.
When starting, I get this:
[+] torch version 1.13.1+cu117 installed. [+] torchvision version 0.14.1+cu117 installed. [!] xformers version 0.0.16rc425 installed. [+] accelerate version 0.17.1 installed. [+] bitsandbytes version 0.35.4 installed. [+] diffusers version 0.14.0 installed. [+] transformers version 4.27.1 installed. Launch errors detected: Incorrect version of xformers installed.
3) Problem with X/Y/Z Plot: for some reason Prompt S/R which was working fine, now it doesn't. I tried to use different tokens, and make sure the syntax is correct. I get this error:
RuntimeError: Prompt S/R did not find prsr:1.0 in prompt or negative prompt.
Or
RuntimeError: Prompt S/R did not find abc:1.0 in prompt or negative prompt.
RuntimeError: Prompt S/R did not find fff: 1.0 in prompt or negative prompt.
I tried all combinations, with space, without, with comma, etc, and the same token in the prompt like "photo of (xyz man: prsr) " or "photo of (xyz man: abc) "
Consideing that Gradio freezes every single time, this is increasing time usage for users on Runpod. I already sort of complained about slow download speed on Runpod's channel, and now we have this to make things even slower.
Our Discord : https://discord.gg/HbqgGaZVmr. How to use custom, different, .safetensors and SD 2.1 on Automatic1111 Web UI. 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 Go...