I worked through all of the steps ytou describe but the process is failing for me at the the VAE Decode Node with the following message: Given groups=1, weight of size [4, 4, 1, 1], expected input[1, 16, 128, 128] to have 4 channels, but got 16 channels instead. Any idea what could be causing this?
Flux Schnell is distilled from Flux Dev. Flux Dev is a higher-quality model but requires more steps, while Schnell usually works in fewer steps, making it faster. The model I recommended in this video is a merge between Schnell and Dev, combining the speed of Schnell with some of the quality of Dev.
How much ram do you have, if around 16 gb then it's worth a try, my one recommendation is to download the lower quantization model, like around q4 for the model, and q3 for the t5 encoder model.
hi there just here to say that was some useful video :) will really improve my stack thanks!
Thanks for this very helpfull tuto ! I saw that you have the same gpu as me 3060 rtx with 12 go which models should you advice me to use ? Q4 ?
I think that q5 is a balanced model, but q4 is not bad either. Ultimately, the choice depends on how much VRAM you want the model to use.
I worked through all of the steps ytou describe but the process is failing for me at the the VAE Decode Node with the following message: Given groups=1, weight of size [4, 4, 1, 1], expected input[1, 16, 128, 128] to have 4 channels, but got 16 channels instead. Any idea what could be causing this?
Do you have flux vae downloaded?
HI...Thanks a lot.....Will it work with GTX 1650 super and 16 GB RAM?
What is the time to generate a image, did you speed up the video?
Yeah video is speeded up, and but I think at lower steps it takes around 30 to 40 seconds on my 3060 12 gb.
@@pixcores what in the hill
Can it work with my specs?
Nvidia 3060 20Gb total memory, 4Gb dedicated memory.
Cpu i7, 32Gb ram
Yeah, i think it should work fine.
Can I run it with 16gb system Ram😢 ihave rtx3060 12gb
I also have the same GPU with the same amount of ram, so yeah you should be able to run this in your pc.
@@pixcores Thanks. Tesed running fine on my computer
how is this different to the Flux schnell? asking by a noob🤥
Flux Schnell is distilled from Flux Dev. Flux Dev is a higher-quality model but requires more steps, while Schnell usually works in fewer steps, making it faster.
The model I recommended in this video is a merge between Schnell and Dev, combining the speed of Schnell with some of the quality of Dev.
@@pixcores what GPU are you using? I’m using RTX 4070 Ti Super 8vram and 16gb memory. Which Uber model and checkpoint model would you recommend.
Will it run in my 4gb graphics, nividia GeForce rtx 3050, i5 12gen???!🥹
How much ram do you have, if around 16 gb then it's worth a try, my one recommendation is to download the lower quantization model, like around q4 for the model, and q3 for the t5 encoder model.
@@pixcores i got only 4gb ram too😭