david/aphrodite-engine: PygmalionAI's large-scale inference engine pygmalion.chat It is designed to serve as the inference endpoint for the PygmalionAI website, and to allow serving the Pygmalion models to a large number of users with blazing fast speeds (thanks to vLLM's Paged Attention). @ 1334a833a400d15e228706138c4f8d7038d45f38

AlpinDale 1334a833a4 feat: AMD ROCm support (#95)		1 anno fa
..
attention	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
quantization	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
activation_kernels.cu	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
cache.h	1aab8a7d6f feat: speedup compilation times by 3x (#130)	1 anno fa
cache_kernels.cu	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
cuda_compat.h	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
cuda_utils.h	1aab8a7d6f feat: speedup compilation times by 3x (#130)	1 anno fa
cuda_utils_kernels.cu	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
dispatch_utils.h	32844c1522 add GELU kernels and remove compile bloat	1 anno fa
layernorm_kernels.cu	7612f33afd feat: fused add RMSNorm kernels (#125)	1 anno fa
ops.h	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
pos_encoding_kernels.cu	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
pybind.cpp	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa
reduction.cuh	1334a833a4 feat: AMD ROCm support (#95)	1 anno fa