All my models are available here
https://www.buymeacoffee.com/razzzp/posts
Disclaimer: This artificial intelligence model ("the Model") is developed solely for artistic expression and technological demonstration, not for creating or distributing illicit or inappropriate content. The creator expressly disapproves of such use. By downloading or using the Model, you ("the User") agree not to misuse it in any manner contrary to these intentions and accept all responsibility for compliance with these terms.
Description
Initial Release
FAQ
Comments (8)
have you tried resizing your loras in kohya_ss?
the quality is great but 800mb is an overkill :)
Damn the result are amazing
Great Results!
What is the best setting for Kohya_ss for Lora training? Like Train batch size, steps, Epoch, etc.
Also on a RTX4090
Unfortunately this Lora is only producing black images for me in vladmandic automatic UI using same basemodel sdxl_sd_xl_base_1.0_0.9vae.
I hope people keep working with SD 1.5 versions for a while (a GOOD while) instead of just jump to XL because it's heavy as hell for low-end machines and wow, 800MB in a LoRa is crazy.
so realistic
I see that you have 1010 regularisation images, can you share where you got them?
Great start! Can't wait to see how this improves. I'm getting better resutls with SD instead of SDXL + this LORA but I think it's just because XL is so new still
Details
Files
scarlet_lora_sdxl_v3-000008.safetensors
Mirrors
94011_scarlet_lora_sdxl_v3-000008.safetensors
scarlt.safetensors
scarlet_lora_sdxl_v3-000008.safetensors
scarlet_lora_sdxl_v3-000008.safetensors
scarlet_lora_sdxl_v3-000008.safetensors
7jo000008.safetensors
Scarlett_Johansson.safetensors
scarlet_lora_sdxl_v3-000008.safetensors
scarlet_lora_sdxl_v3-000008.safetensors
Available On (2 platforms)
Same model published on other platforms. May have additional downloads or version variants.













