Recent content by kharo88

I agree that the guided one looks a lot better. Good job!

It's funny that the first one on your image there always has good bounce/physics but the rest not so much. :) And I guess the settings were the same for all of them. Things like this would probably drive me crazy if I wanted to do something like your JAV trailer.

Yeah, unfortunately, I lost heart, but sometimes I feel tempted to give training a LoRA for the 5B model a try, AI says it might work. Just might... And yes, I also used the GGUF versions (Q4_K_M for both high and low), but not the lightning LoRAs (though AI mentioned those as a possibility)...

Oh, I don't have it anymore, I have already uninstalled ComfyUI and got rid of everything, but it was just the basic wan 2.2 14B I2V workflow, nothing extra was added.

I think the first one might be Shinoda Ayumi.

Yeah, the base video is 854 x 480 and low quality; AI told me to scale up the screenshots, so that's what I did. And you're right, it's far from ideal, but that's the best that I could do. I would have been curious to see how closely a LoRA for wan 2.2 would have resembled the girl on the...

Here are the two videos created based on a different image from my first try at this if you are curious: https://limewire.com/d/QkcH7#RueYqDF8Xk I tried to recreate your success by copying your image and prompt, but now ComfyUI just gets stuck at the first KSampler. I've been trying to...

I think I'm done. I have made two LoRAs, neither worked, then I tweaked it, and now it would take forever to train one, so I put that aside for now to get wan 2.2 14B running. I eventually got it to run, chose one of the screenshots that I wanted to train a LoRA with, but the videos were not at...

Yes, you're correct, my goal is to train a character LoRA to use with, say, wan 2.2 14B. I'm not familiar with any image models, but I'll look into that. I didn't even think about using an image gen AI, then using the resulting pictures with a video gen AI, but yeah, you're definitely right that...

I'm not sure, Casshern. I think I could train a LoRA for wan 2.2 5B, but compared to the 14B version it just looks much worse, so I'd like to train a LoRA for use with the 14B model and ChatGPT basically says that "you can give it a try, but don't be surprised if you get an out of memory crash"...

Unfortunately I'm still shafted. I did upgrade my PC, so now instead of 6 GB VRAM I've got 16 GB, and instead of 8 GB of system RAM I've got 32 GB, and this would be enough to run an AI model, but the thing that I didn't know beforehand is that this is not really enough to train LoRAs, and...

Does anybody who speaks Japanese can tell or find out if the Japanese audience is okay with this? They are the only ones who could demand the industry go back to real cumshots. I'm not sure if the studios would listen, but it would be nice to know the opinion of the Japanese customers at least.

Wow! This is pretty good!

Well, my GPU only has 6 GB of VRAM and I've only got 8 GB of RAM, so I suppose my dreams will just have to be dreams for now. But thank you, it's good to know that what I have in mind is possible. Hopefully next year I'll be able to upgrade my rig. Until then I'll keep an eye on this thread!

I can't really share as the only time that I tried to make something with LTX (the online version) it didn't come out anything like what the prompt was. In fact, it seemed like the online version didn't even have adult content enabled, so I'd have to DL it. My PC is very weak though, and taking...