[N] Introducing Stable Fast: An ultra lightweight inference optimization library for HuggingFace Diffusers on NVIDIA GPUs

# What is this?

[stable-fast](https://github.com/chengzeyi/stable-fast) is an ultra lightweight inference optimization library for __HuggingFace Diffusers__ on __NVIDIA GPUs__.
`stable-fast` provides super fast inference optimization by utilizing some key techniques and features:

- __CUDNN Convolution Fusion__: `stable-fast` implements a series of fully-functional and fully-compatible CUDNN convolution fusion operators for all kinds of combinations of `Conv + Bias + Add + Act` computation patterns.
- __Low Precision & Fused GEMM__: `stable-fast` implements a series of fused GEMM operators that compute with `fp16` …

bias convolution cudnn diffusers features functional fusion gpus huggingface inference library machinelearning nvidia nvidia gpus operators optimization series