AlpinDale d9b65e6c5f feat: DeepSeek MoE support (#237) 1 年之前
..
all_reduce 6305e6f3f2 fix: no repeated IPC registration (#227) 1 年之前
attention 31c95011a6 feat: FP8 E5M2 KV Cache (#226) 1 年之前
hadamard 5d288aa76c feat: add fast hadamard transformation kernels (#232) 1 年之前
moe d9b65e6c5f feat: DeepSeek MoE support (#237) 1 年之前
punica c0aac15421 feat: S-LoRA support (#222) 1 年之前
quantization aebd68c632 feat: backport kernels (#235) 1 年之前
activation_kernels.cu 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 1 年之前
cache.h 31c95011a6 feat: FP8 E5M2 KV Cache (#226) 1 年之前
cache_kernels.cu 31c95011a6 feat: FP8 E5M2 KV Cache (#226) 1 年之前
cuda_compat.h 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 1 年之前
cuda_utils.h 31c95011a6 feat: FP8 E5M2 KV Cache (#226) 1 年之前
cuda_utils_kernels.cu 31c95011a6 feat: FP8 E5M2 KV Cache (#226) 1 年之前
dispatch_utils.h d9b65e6c5f feat: DeepSeek MoE support (#237) 1 年之前
layernorm_kernels.cu 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 1 年之前
ops.h d9b65e6c5f feat: DeepSeek MoE support (#237) 1 年之前
pos_encoding_kernels.cu 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 1 年之前
pybind.cpp d9b65e6c5f feat: DeepSeek MoE support (#237) 1 年之前
reduction.cuh 2755a48d51 merge dev branch into main (#153) 1 年之前