search

s9roll7 / animatediff-cli-prompt-travel

animatediff prompt travel
Apache License 2.0
1.19k stars 104 forks source link

Support for absolute paths? #181

Open tnil25 opened 1 year ago

tnil25 commented 1 year ago

I've created a desktop UI for animatediff-cli-prompt-travel that people are really enjoying: https://civitai.com/models/192131/diffex-desktop-ui-for-animatediff-prompt-travel

However the one request I keep getting is the ability for the user to use their models from an existing stable diffusion installation (like A1111 or ComfyUI) so they don't have to make copies of all their models. I know ComfyUI supports this by specifying the path to A1111.

Is there any possibility this could be implemented in some way?

s9roll7 commented 1 year ago

I am using symbolic link https://learn.microsoft.com/en-us/windows-server/administration/windows-commands/mklink

s9roll7 commented 1 year ago

I'll make it configurable in the config file eventually.

tnil25 commented 1 year ago

Awesome, thank you!

Support for SDXL turbo?

s9roll7 commented 12 months ago

Awesome, thank you!

Support for SDXL turbo?

https://civitai.com/models/215485?modelVersionId=242807

tnil25 commented 12 months ago

Gonna try this. I assume you use it like any other LoRA in cli?

s9roll7 commented 12 months ago

Gonna try this. I assume you use it like any other LoRA in cli?

yes

alanhuang67 commented 12 months ago

Hello, follow the question, when I use sdxl-turbo lora,

  1. do I need to use lcm setting?
  2. do I need to use the sampler "LCM"? or just use orgional sampler？
  3. anything sitting we need to pay attention to? Thanks!
s9roll7 commented 12 months ago

Looks like I need to make the lcm sampler selectable.

s9roll7 commented 12 months ago

Hello, follow the question, when I use sdxl-turbo lora,

  1. do I need to use lcm setting?
  2. do I need to use the sampler "LCM"? or just use orgional sampler？
  3. anything sitting we need to pay attention to? Thanks!

https://github.com/s9roll7/animatediff-cli-prompt-travel/assets/118420657/b148ff66-ae32-4d6b-8cc0-538f89a3b81b

prompt.json

animatediff generate -c config/prompts/sample_turbo_sdxl_test.json -L 80 -W 768 -H 1280 sd_xl_turbo_lora_v1.safetensors": 0.75 8(+1)steps generation time 122 sec

I think the output is cleaner than lcm lora

alanhuang67 commented 12 months ago

looks grear!

I saw the new config support gradual_latent_hires_fix, does this support img2img?

s9roll7 commented 12 months ago

looks grear!

I saw the new config support gradual_latent_hires_fix, does this support img2img?

It should work, but I can't remember if I debugged it or not.