hyppyhyppo
hyppyhyppo
It is better explained in the Wiki, check the concept, embedding and FAQ pages, you also can jump in the Discord server for further questions. Just that `` works as...
Not sure to fully catch your request : " I would love it if it would be possible to have a selection add from txt file where we can select...
I mean one concept with generated prompts and option "From text file per sample" and another with the option "from single text file" where you select your unique text file....
Well, honnestly I still don't really catch the request but I still think some workaround can be found with the existing options. You can use several options for a concept...
I had once a similar error, check if this article can solve it. https://nvidia.custhelp.com/app/answers/detail/a_id/5490
Just for info, it also happens for Flux Lora, I didn't test with other models. Different rank/alpha, maybe layers too.
Hi, I'm probably one of the rare Windows user having 2 GPU and can run some tests. first tests managed to solve https://github.com/Nerogar/OneTrainer/pull/816#issuecomment-3195656013 . that was due because some nvidia-smi...
Sharing my tests on Windows, SDXL fine tune, default settings, 163 images all cropped to the training res (1024x1024 and 832x1216), local batch 4, accumulation step 1 then 2, 2...
An SDXL support would be great indeed. I used it a lot for merging embeddings together or with additional tokens.
Note on my latest test from yesterday. I noticed a UI Bug where I could not change the model output destination.