AlpinDale e31c6f0b45 feat: refactor modeling logic and support more models (#274) 10 hónapja
..
all_reduce 6305e6f3f2 fix: no repeated IPC registration (#227) 11 hónapja
attention 8e1cd54497 fix: do not include fp8 for rocm (#271) 10 hónapja
hadamard 5d288aa76c feat: add fast hadamard transformation kernels (#232) 11 hónapja
moe 7d6ba53602 feat: fused top-k kernels for MoE (#273) 10 hónapja
punica c0aac15421 feat: S-LoRA support (#222) 11 hónapja
quantization 8e1cd54497 fix: do not include fp8 for rocm (#271) 10 hónapja
activation_kernels.cu e31c6f0b45 feat: refactor modeling logic and support more models (#274) 10 hónapja
cache.h 31c95011a6 feat: FP8 E5M2 KV Cache (#226) 11 hónapja
cache_kernels.cu 8e1cd54497 fix: do not include fp8 for rocm (#271) 10 hónapja
cuda_compat.h 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 11 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 d9b65e6c5f feat: DeepSeek MoE support (#237) 11 hónapja
layernorm_kernels.cu 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 11 hónapja
ops.h e31c6f0b45 feat: refactor modeling logic and support more models (#274) 10 hónapja
pos_encoding_kernels.cu 8fa608aeb7 feat: replace Ray with NCCL for control plane comms (#221) 11 hónapja
pybind.cpp e31c6f0b45 feat: refactor modeling logic and support more models (#274) 10 hónapja
reduction.cuh 2755a48d51 merge dev branch into main (#153) 1 éve