Steward Garcia

@JustMaier I contributed to implementing and improving hardware acceleration support in CUDA [#75](https://github.com/leejet/stable-diffusion.cpp/pull/75). I can say there is still plenty of room for improvement, but I was experimenting with ways...

>it doesn’t use tensor cores. @rmatif Looking at the cuDNN code you wrote, I see that it is indeed configured to use the tensor cores with `cudnnMathType_t math_type = (i...

Honestly, I never noticed any improvement with ggml’s Flash Attention kernels. In terms of memory, it’s noticeable (in SD 1.5 it’s only about -1MB from what I’ve seen), but in...

The creator of this project himself tried to implement a video generator, Stable Diffusion Video, but it didn’t work for him and he gave up. Implementing a video generator is...

It is normal since SD 1.5 does not support resolutions that are not multiples of 2, and as long as one side of the image is 512, there is no...

It seems that it is only available for matrix multiplications and not for custom operations. Apologies for my ignorance.

I also intended to use that stable diffusion project with flame.cpp and create something like ChatGPT + DALLE-3, but the truth is that the prompt generation quality of the models...

Does this only work for Expo, or also in vanilla React Native?

It would be great, this library has a lot of potential. The [simple](https://github.com/ggerganov/ggml/tree/master/examples/simple) example is a good starting point to learn the basics of GGML and how to perform a...