GitLab

Issue/200/fix: Updata FP16 and FP32 Tolerance for Add and SwiGLU

修复昇腾调用aclSetAclOpExecutorRepeatable的潜在危险

issue/131: rearrange 算子 - CUDA

同时修复 CPU rearrange 算子内存泄漏问题

issue/172 Add 算子 CPU & CUDA

issue/194/fix: Fix ElementwiseInfo Meta Allocation

issue/197/fix  修复rope cuda算子中dimx使用uint32_t代替unsigned int

issue/194: fix meta mem size allocation and change cpu DeviceImpl::create to return INFINI_STATUS_NOT_IMPLEMENTED

issue/189/docs: update README.md

Signed-off-by: YdrMaster <ydrml@hotmail.com>

Revert "添加swiglu算子测例"

添加swiglu算子测例

issue/8: causalsoftmax算子-昇腾

Issue/48 Rope CPU & CUDA

使用maskedInplaceOp和SoftmaxOp实现CausalSoftmaxOp

Issue/127/feat. General Elementwise Framework with Refactored SwiGLU (CPU & CUDA)

issue/127: change meta within ElementwiseInfo to std::vector<size_t> for correct alignment and change the reference name of the Opaque struct to Opaque instead of struct Opaque

issue/127: Add arguments to CREATE_ELEMENTWISE_PLATFORM_DESCRIPTOR macros for indirecting variable names, change DeviceImpl to use Result for the return type of the create function, change CEIL_DIV