images used for training
hi!
I'm not sure if this is the correct place to ask, but would it be possible to see all the images you used for training for this data set, I really want to put myself into elden ring, just a screen shot of the directory including the images will be fine.
its okay if this cannot be done though
Sure thing, her you go:
omg thankyou so much!!
@nitrosocke
Amazing work! I'm loving your models and having a ton of fun with the Archer model <3
What does the model training process look like? Can you point me to the right direction?
Thank you, glad you enjoy these models so much!
There are a few videos about it on YouTube that should get you started.
I use the Shivam repo for my trainings, if you want to get the same results. Another popular choice is the one by JoePenna.
Hi, your results are gorgeous! Do you tweak anything special in Shivam's DreamBooth_Stable_Diffusion notebook besides changing the number of steps?
Thanks a lot
These where my settings for training:
prior_loss_weight= 1.0
train_batch_size= 1
gradient_accumulation_steps= 1
learning_rate= 1e-6
lr_scheduler= "constant"
max_train_steps= 3000
train_text_encoder
gradient_checkpointing
Hey! Can you give the link of Shivam's DreamBooth_Stable_Diffusion and what are your computer settings
Hey! Can you give the link of Shivam's DreamBooth_Stable_Diffusion
https://github.com/ShivamShrirao/diffusers/
And you should find more than what you asked for in Nitro's guide on it : https://github.com/nitrosocke/dreambooth-training-guide/blob/main/README.md
What sort of magic is this? Only 23 images, and all those wonderful images coming out. Would it make sense to add more images?