GitLab

Issue/563 Add metax support for topkrouter

issue/812: 增加CUDA_ARCH选项与CUTLASS支持

issue/798 - fix operator device handling

issue/792 - add moore threads device sync to tests

Issue/794 - Use native musaEventElapsedTime API.

issue/759 - list input adjustments

issue/757 - support equal_nan in test debug

issue/753: kunlun gelu kernel

Issue/744 kunlun softplus

issue/744: softplus

issue/744: kunlun softplus

issue/744: delete F64

issue/723 - Expose test framework to other packages

issue/716: Add save feature for existing test cases

issue/746: 修复causal_softmax在长宽在1024边缘的计算错误

Signed-off-by: Ceng23333 <441651826@qq.com>

* issue/739 - support batched RoPE on Nvidia and CPU

* issue/739 - metax, moore batched rope

* issue/739 - adjust metax flags

* issue/739 - added a rope module interface to forward inplace in output tensor

issue/741  暴露 infinicore_cpp_api 类继承

issue/722 - adjusted cuda rearrange for shape (8, 4, 20, 64)

issue/713 - 为c++添加 RowParallelLinear 和 ColumnParallelLinear

issue/730 - optimized debug

 Module 支持单个张量加载

Issue/714 - rearrange和random_sample测试修改