Is there any way to resume LoRa training after the overnight automatic Windows update that restarted
Is there any way to resume LoRa training after the overnight automatic Windows update that restarted my PC?
LHi all - followed the Flux training Lora guide/tutorial. Does my speeds seem correct for a 3090? It’s roughly 9 it/s using the level 2 workflow.
I keep seeing references to a “massive speed increase” with torch 2.5, but the 1 click installer has I believe installs 2.7? So I imagine that’s already implemented.
FluxGym is substantially faster, around 2 it/s, if memory serves me. But as I am just learning about all this, is a bit difficult for me to compare apples to apples/settings to settings. FluxGym while easy to use, doesn’t have the most friendly of interfaces when digging into additional settings.
I keep seeing references to a “massive speed increase” with torch 2.5, but the 1 click installer has I believe installs 2.7? So I imagine that’s already implemented.
FluxGym is substantially faster, around 2 it/s, if memory serves me. But as I am just learning about all this, is a bit difficult for me to compare apples to apples/settings to settings. FluxGym while easy to use, doesn’t have the most friendly of interfaces when digging into additional settings.
!how could i fix this?
F
Fyes already implemented
FFluxGym is not faster using different config
Fand a weaker one
Fmodel download looks like failed / missing
Fhugging face is down atm
!ah okay
!thanks for the reply, i've managed to fix it but now i run into this. Can this been the file is corrupted?
Fyep can be corrupted
LThat’s what I figured. Appreciate the response!
Fyou are welcome
Fserializing messages usually related to model corruption
JScammer
Fyep banned
PHi! I have a question, maybe someone can help me.
Can we retrain a fine-tuned model to create another fine-tuned model based on it instead of starting from Flux1dev?
Thanks!
Can we retrain a fine-tuned model to create another fine-tuned model based on it instead of starting from Flux1dev?
Thanks!
SWhat’s the best program or extension to use right now to make a sd model of myself
SI will take any recommendations
Fyes but it gets worse as you train also if same concept it will bleed
Fonetrainer
FYouTubeSECourses
In this tutorial, I am going to show you how to install OneTrainer from scratch on your computer and do a Stable Diffusion SDXL (Full Fine-Tuning 10.3 GB VRAM) and SD 1.5 (Full Fine-Tuning 7GB VRAM) based models training on your computer and also do the same training on a very cheap cloud machine from MassedCompute if you don't have such compute...
�I trained a flux lora on cicitai, and for testing i went with a real person. On generations, skin gets plastic and shiny...not like my dataset. Using heun beta 30 steps. Curious about what might cause this
Fcivitai training weak
Fsee my latest tutorial i shown myself trained result
FYouTubeSECourses
FusionX: The BEST AI Video Model? + FLUX Hyper-Realistic Upscaling (One-Click Setup!). Struggling to create high-quality AI videos and hyper-realistic images? This tutorial is your ultimate solution! I'm introducing the incredible new Wan 2.1 FusionX model and a game-changing 2x latent upscaler for the FLUX model, all made incredibly simple with...
F13:48 Generating a High-Quality Image with The Official FLUX Preset
14:50 Using Automatic Face Segmentation & Inpainting with FLUX
16:05 The Ultimate Upgrade: Applying The FLUX 2x Latent Upscaler Preset
16:32 Final Result: Comparing Standard vs. 2x Upscaled Image Quality
16:50 Outro & Sneak Peek of The New Ultimate Video Processing App
14:50 Using Automatic Face Segmentation & Inpainting with FLUX
16:05 The Ultimate Upgrade: Applying The FLUX 2x Latent Upscaler Preset
16:32 Final Result: Comparing Standard vs. 2x Upscaled Image Quality
16:50 Outro & Sneak Peek of The New Ultimate Video Processing App
MScammer
THi
TUsing a different model than Flux with my Lora looses the likeness. Is it possible to train a LoRA based on a Flux Model other than the Flux itself ?
Fyes any flux variant
Fyou can train
Tshould I check and verify the weigths ect as in the 1st place mixing my lora with that model is altering my model details or it'll be done automatically?
Fwell it will be different
Fif you have trained on another base model
Fbut probably with flux i assume it would work decent
Fsince flux fine tunes are not so much different
Fusually
LHave any examples to share? If not that’s fine. Just curious how “plastic” it is. Plastic type skin is pretty inherent with Flux regardless. I have read there are some Loras that can help, I am testing way too many different things at the moment so can’t suggest.
And everything I have read training to Flux 1 Dev is by far the best.
What other models are you training when you say it looses its likeness?
And everything I have read training to Flux 1 Dev is by far the best.
What other models are you training when you say it looses its likeness?
PDr Furkan, I saw you released an update for SDXL finetunning today, 20/june. Are you planning any update for FLUX finetunning or your last release for FLUX are the best yet?
Fi recently tested flux fine tuning with more than 20 experiments
Fnone yielded better 
PDr using your tutorial and trainning a FLUX finetunning in Massed Compute, about choosing the GPU, is there a better GPU to choose? For better I'm only considering faster trainning and the relation cost x benefit, I know that the prices increase. I don't understand one thing, if I choose a H100 x8, or a 8 H100 boards, is it eight times faster than an 1 H100? or it doesn't work this way? Can you tell me what is the best board option in Massed Compute, ignoring price, for a 100 images finetunning of a person?
Ffor lora it will be almost 8 x
Fbut for dreambooth last time i tested it wasnt working properly
Fbest price performance is RTX a6000 ADA i think
PIn your post, about Dreambooth finetunning, you said: "...single RTX 4090 is almost same speed as RTX A6000 for Fine Tuning." Does it mean that if I have a RTX 4090 can I train, locally, at the same speed as a RTX a6000 that has 48gb of vram, double vram than mine 4090?
F4090 is faster when fits into VRAM. current not fitting but with block swapping it matches almost
Fwhat is your second / it ?