Nice and Big Thank!!

@goshniiAI

24 күн бұрын

thank you for stopping by.

i like your channel, thanks bro

@goshniiAI

19 күн бұрын

i appreciate you

Nice!! Thanx!! I tried it and it works great. I have a question though.... is there any way to make the transitions and the overall video longer? I changed the batch from 96 to 120 but didn't notice a difference between the time for each image nor the overall video length. Thanx!

@goshniiAI

Ай бұрын

Yes, that is a strong workflow and you're correct. I am aware that using the default 96 frames produces the greatest outcomes from the workflow. unless you are certain of what you are doing, you may encounter some difficulties.

Excuse me, master, is it possible to make a short animation in sequence with the reference images or is there a better method for that?

@goshniiAI

14 күн бұрын

Yes, that is achievable but the workflow now limits us to only 64 frames of time, so I believe this could be done in batch sequences. However, different workflows could make things much easier.

@itanrandel4552

14 күн бұрын

@@goshniiAI ty,for you time master

is this workflow works on animatediffXL or hsxl? I have tried, but failed

@goshniiAI

Ай бұрын

The workflow performs best with SD1.5 models. If SDXL or Hotshot failed, you could use the workflow's upscale nodes.

when i press prompt queue im only getting black boxes at the bottom

@goshniiAI

Ай бұрын

If the black boxes are far below the canvas of comfyui, simply click cancel next to them. additionally Check that all of your nodes are up to date. A version mismatch may cause this kind of problem.

Depressing to see these still focusing on 1.5 models and sdxl. The difference in prompt cohesion and image quality is night and day between 1.5 and sdxl. Sadly all the time content creators focus on 1.5. I thank you for the video, and i know the time that goes into these, but please consider doing one for sdxl.

@goshniiAI

29 күн бұрын

Thank you for your thoughtful feedback! The advancements in SDXL are indeed remarkable, and I appreciate your suggestion. I'll definitely consider creating SDXL-related videos to help everyone get the most out of its possibilities.

@GamingDaveUK

29 күн бұрын

@@goshniiAI Appreciated. At the minute 1.5 is on a bit of a feedback loop: 1) People are using 1.5 more than sdxl as there are more guides for it. 2) Content creators focus on 1.5 as more people are using it, so they make guides for 1.5. 3) tool devs see that 1.5 is being used more than 1.5 so focus energies on 1.5. 4) and were back to 1) When we finally get SD3 I am hoping it breaks this loop.... but I must admit, I doubt it.

@sudabadri7051

22 күн бұрын

You can try change this to hotshotxl animate diff and and ipiv workflow perhaps, while sdxl has lots of benefits. Pose recognition is not one of them sd1.5 still is viable for a few things and since its 512x512 it has lots of cool research models because training is easier at that res. Guides are also a part of that cycle.

@goshniiAI

21 күн бұрын

@@sudabadri7051 entirely correct that there are advantages to each kind of stable diffusion. While SDXL has many advantages, SD1.5 is still more enjoyable, particularly for certain purposes such as pose recognition and 512x512 resolution training. The guide cycle you mentioned is spot-on, too.

@goshniiAI

21 күн бұрын

@@GamingDaveUK I’ll do my best to create content for both versions to help bridge the gap, It's a difficult cycle to break, but the chance of SD3 is interesting. hopefully the focus will be shifted and encourage more exploration with SDXL.