Started empirically experimenting for finding the best training hyper parameters for Stable Diffusion 1.5 based models
Added 2024-01-25 00:00:56 +0000 UTC- Using Kohya SS GUI as trainer script.
- Rented 8 GPU machine from RunPod and already doing 8 different trainings right now.
- If you have any input / feedback / suggestion let me know please.
Comments
I trained both on 768 and 1024. 768 works best and you can generate very good images at 1024. I will post full config. before testing OneTrainer. OneTrainer uses only 10300 MB VRAM on Windows
Furkan Gözükara
2024-01-26 22:39:09 +0000 UTCCan you share the base config you are using? Are you training at 512x512 resolution or higher?
mypatreonemailacc
2024-01-26 15:54:04 +0000 UTCoptimizer, optimizer arguments, learning rate. still testing heavily
Furkan Gözükara
2024-01-25 15:35:04 +0000 UTCVery interesting. What are the main parameters you between the trainings? I.e. which training parameters in your opinion makes the biggest difference in quality?
chriseppe
2024-01-25 14:28:07 +0000 UTC
