Type | |
Stats | 127 |
Reviews | (24) |
Published | Oct 19, 2024 |
Base Model | |
Training | Steps: 1,000 Epochs: 1 |
Usage Tips | Clip Skip: 1 Strength: 1 |
Hash | AutoV2 A0D984866C |
This is a LORA model of Not SNSD - Taeyeon Flux1.D. It should work well with flux1-dev-fp8.safetensors.
Higher Rep can be found here :
https://ko-fi.com/tissue_ai/shop
https://www.patreon.com/Tissue_AI
Using Forge-UI : https://github.com/lllyasviel/stable-diffusion-webui-forge
Tested settings :
Diffusion in Low Bits : Automatic (fp16 LoRA)
Lora weight : 0.8-1
Sampling Steps : 20-30
Distilled CFG : 3.5-5
Sampling Method : DEIS / Euler
Schedule Type : Simple
Hi-res fix HIGHLY Recommended
Hi-Res Sampling Method : LMS
Checkpoint :
flux1-dev-fp8 : https://huggingface.co/lllyasviel/flux1_dev/blob/main/flux1-dev-fp8.safetensors
For all these image Prompt u can check my discord : https://discord.gg/9fKcrWzagr
Download Here for Forge-UI :
https://github.com/lllyasviel/stable-diffusion-webui-forge/releases/download/latest/webui_forge_cu121_torch231.7z
Download Here for Vae/Text Encoder :
Vae : https://huggingface.co/black-forest-labs/FLUX.1-dev/resolve/main/ae.safetensors?download=true
Clip_l :
https://huggingface.co/comfyanonymous/flux_text_encoders/resolve/main/clip_l.safetensors?download=true
t5xxl_fp8 : https://huggingface.co/comfyanonymous/flux_text_encoders/resolve/main/t5xxl_fp8_e4m3fn.safetensors?download=true