Heliumrich

It seems LoRAs don't unload when I bypass the load lora node. I need to manually unload the model (comfyui-manager button) to be able to unload any lora. Otherwise if...

Hmm, I checked logs (by running `MOZ_LOG="FFmpegVideo:5" firefox`) and I saw lots of these when playing a YT video: `[RDD 14015: MediaPDecoder #1]: D/FFmpegVideo FFVPX: Frame decode takes 0.32 ms...

> Something that is helpful for me when debugging the drivers is running the following command in the terminal while playing. a video: > > `watch -n 2 nvidia-smi `...

> This blocklist reason is perfectly normal and it's just the new name for the old `FEATURE_FAILURE_VAAPI_NO_LINUX_NVIDIA` that existed ever since firefox added hardware acceleration. I guess that back in...

Ok so the only way to have this work as expected, is to have "borderless" and "fullscreen borderless" in "window management" UNTICKED. Then set the game to borderless in its...

the 5B model is already pretty small (and pretty bad tbh) so I don't really understand the request. the 14B T2V and I2V ones would be way more useful

> I have low vram and there is some block swapping when resolution or num frames are set as I need them. This makes the quantization desirable so the model...

I hope support for Wan-animate will come too. Look amazing and even more useful than base wan (for me) https://github.com/nunchaku-tech/nunchaku/issues/710

@lmxyy Is this your next model support ? Or you're waiting for Wan 2.5 (though I doubt it will be open-sourced) Please support a video model before adding support for...

Sad, Wan2.2 can take like 3minutes on a RTX4080 16GB (fp8 is too big for 5sec, I have to use slow GGUF Q6_K_M)