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). @ chore/test-updates

AlpinDale 40f63268ee disable new layernorm kernels for CUDA < 12.0		9 hónapja
..
all_reduce	29eaded422 fix and re-enable custom all-reduce	9 hónapja
attention	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
backup	f8dfac6372 chore: attention refactor and upstream sync apr01 (#365)	9 hónapja
cpu	fb23720c72 fix CPU build	9 hónapja
hadamard	5d288aa76c feat: add fast hadamard transformation kernels (#232)	11 hónapja
moe	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
punica	c0aac15421 feat: S-LoRA support (#222)	11 hónapja
quantization	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
activation_kernels.cu	3d6695cfbb feat: add approximate gelu activation kernels (#370)	9 hónapja
cache.h	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
cache_kernels.cu	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
cuda_compat.h	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
cuda_utils.h	31c95011a6 feat: FP8 E5M2 KV Cache (#226)	11 hónapja
cuda_utils_kernels.cu	31c95011a6 feat: FP8 E5M2 KV Cache (#226)	11 hónapja
dispatch_utils.h	f8dfac6372 chore: attention refactor and upstream sync apr01 (#365)	9 hónapja
layernorm_kernels.cu	40f63268ee disable new layernorm kernels for CUDA < 12.0	9 hónapja
ops.h	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
pos_encoding_kernels.cu	e702f587cf feat: add batched RoPE kernels (#371)	9 hónapja
pybind.cpp	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja
reduction.cuh	071269e406 feat: FP8 E4M3 KV Cache (#405)	9 hónapja