[10/n] Migrate cuda_view and silu_and_mul_per_block_quant kernels to torch stale ABI. (#44334)
This commit is contained in:
@@ -18,7 +18,7 @@
|
||||
#include <torch/csrc/stable/tensor.h>
|
||||
#include "libtorch_stable/torch_utils.h"
|
||||
#include "cutlass_extensions/torch_utils.hpp"
|
||||
#include "cutlass_extensions/common.hpp"
|
||||
#include "libtorch_stable/cutlass_extensions/common.hpp"
|
||||
|
||||
#include "get_group_starts.cuh"
|
||||
#include "cutlass_extensions/epilogue/scaled_mm_epilogues_c3x.hpp"
|
||||
|
||||
@@ -21,7 +21,7 @@
|
||||
#include "cutlass/util/packed_stride.hpp"
|
||||
#include "cutlass/util/mixed_dtype_utils.hpp"
|
||||
|
||||
#include "cutlass_extensions/common.hpp"
|
||||
#include "libtorch_stable/cutlass_extensions/common.hpp"
|
||||
#include "cutlass_extensions/epilogue/scaled_mm_epilogues_c3x.hpp"
|
||||
|
||||
#include <cuda_runtime.h>
|
||||
|
||||
Reference in New Issue
Block a user