vans554

Pretty interesting. Using nightly + cu121 im getting 8.18 it/s, another 5% improvement vs 7.78 that cu118 gave.

I accidentally stumbled upon this and did not expect such a speedup. It seems anything less than cu118 does not properly support the RTX4090 (or H100).

Bumping to cuda12.2 with pytorch2.0.1+cu118 made my SDXL go 50% faster and ESRGAN 80% faster on the 4090.

I put Karlo diffusion by Kakao running as an inference endpoint on GPUX. Karlo Diffusion is more controllable, similar to Dalle-2.

Make some images here https://explorer.gpux.ai/inference/gpux/karlo

Thanks.

Yea but this is irrelevent unless you implying that FILM is better quality while nvfruc is lower quality but faster so applicable for realtime?

if nvfruc is better quality + faster then the comparison is that nvfruc is better.

very cool how does this compare to NVIDIA nvfruc?

https://docs.nvidia.com/video-technologies/optical-flow-sdk/...

People still go to traditional-school?

Much needed thank you for doing gods work.

DDD is just microservices without head of line blocking not sure why everyone hates it. Instead of having 5 microservices you just build an umbrella app exposing the same 5 API. Calls into the umbrella are now at microsecond speed vs 10s of milliseconds in most (run of the mill publicly available) clouds.

Mouse lag needs to be fixed.

Would be cool to move realworld assets into virtual worlds, and vice versa.