🟩 "Advanced All In One TEA ☕

This workflows includes settings and values I use more often after daily use since day1.

It's based on this particular tea cache nodes

https://github.com/TTPlanetPig/Comfyui_TTP_Toolset

with 3 methods to upscale faster,

plus a lightweight captioning system for I2V and V2V, that consume only additional 100mb vram.

Upscaling can be done in three ways:

1.The first method is by enabling latent upscale + Refine. This is my favorite and also the fastest.

2.The second method is directly upscaling using the model (Refine is optional)

3.The third use a video upscaler (will add some details and contrast) + Refine

Three different methods, more choices based on preferences.

Requirements:

-Tea cache nodes

-ClipVitLargePatch14

-RealESR General x4 v3

Bonus TIPS:

Here an article with all tips and trick i'm writing as i test this model:

https://civitai.com/articles/9584