Hello Dr. Furkan, I tired literally everything, reinstalled everything, made a whole new pod, tried
Hello Dr. Furkan, I tired literally everything, reinstalled everything, made a whole new pod, tried mimicpc instead of runpod, same thing
CDo you know why using a LoRA can cause this drop in quality? Generated with ComfyUI.
Apoor dataset quality
CNah it’s something else, fine tuning on this dataset does not produce such artefacts.
Tchatgpt made me this graph from your raw testing data
Ksame
Kno CLIP/text encoder weights in checkpoint, the text encoder model will not be loaded.
CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
clip missing: ['text_projection.weight']
CLIP/text encoder model load device: cuda:0, offload device: cpu, current: cpu, dtype: torch.float16
clip missing: ['text_projection.weight']
KEven though my Clip files are installed, I'm getting a 'clip missing' error. Has anyone experienced this before, or do you know the solution?
Agot an answer on swarm server.
Go on the left side of swarm on the search bar "Filter parameters" and search for the keyword preview. There's a tik option on the Swarm Internal tab that reads "No previews" make sure to uncheck it
Go on the left side of swarm on the search bar "Filter parameters" and search for the keyword preview. There's a tik option on the Swarm Internal tab that reads "No previews" make sure to uncheck it
Tsounds amazing, much better than F5-TTS
Acheck your DMS pls
Ayeah but poor language support
TF5-TTS had only 2 languages, zonos have 5
AF5 could be finetuned tho
Mwhats the optimal amount of steps to have for flux lora training? i have 175 images and 180 reg images
Mbtw training is very slow, i am on 48 vram, how can i make it faster?
KWhere is the unified download for swarmui again?
Kwill try ithank u
Kanyone else getting this on windows
Kim using my SSD and still have this issue with swarm ui
K
K@Furkan Gözükara SECourses this is using the unified downloader, i ran sawrm ui with launch-windows.bat
Kright before it crashes i see this
Ndo you get the same error if you try a different model?
Fye i am following on my list
Fthis happens when using lora but fine tuning fixes
Fif you are gold member i can connect and check your setup
Fgetting faster gpu
Fand bigger batch size
FPatreon
Get more from SECourses: FLUX, Tutorials, Guides, Resources, Training, Scripts on Patreon
F16 gb ram your issue
KYeah i noticed 
Ki guess i should just upgrade RAM inside my PC
F100%
Fgo to 2x48 gb if your mobo supports
For 2x32
K2 ram sticks u mena?
Kmean*
Fyes
K@Furkan Gözükara SECourses
Fi dont know from memory i use swarmui 
Ji do think Timson have the right answer, i had similar results with overtrained Loras, but in the end, it is mostly the resolution.
Mtrained a LoRA for flux exactly like the tutorial (for human images), the results are nothing like the trained pictures, and its not even a consistent character, just keeps pulling out new character as if i am not using any lora
Mand even tho i used hundreds of images, the size of each of the lora files was only 18MB
Fstripes happens when you upscale with lora
Ffine tuning doesnt have that issue
Fnot our config 
Ddoes the sort by similiarity app run on the gpu or the cpu? mine is using 93% of my cpu but 3% of my gpu?
Fit uses gpu
Fbut you have to follow requirements