Age | Commit message (Expand) | Author |
---|---|---|
2021-08-20 | Rename [Cl|Cpu]GemmConvolution to [Cl|Gpu]GemmConv2d | Georgios Pinitas |
2021-08-13 | Ensure correct transformed matrices are used in CpuGemmConvolution | Georgios Pinitas |
2021-07-29 | Port NEConvolutionLayer | Michalis Spyrou |
2021-07-23 | Avoid allocation of auxiliary memory in CpuGemmConvolution | Georgios Pinitas |
2021-07-23 | Pass fast math flag for correct GEMM3D validation support | Georgios Pinitas |
2021-07-22 | Expose fast_math mode for GEMM through BFloat16 | Georgios Pinitas |
2021-07-22 | Update GEMM assembly kernels | Georgios Pinitas |
2021-07-15 | Port NEGEMMConvolutionLayer | Manuel Bottini |