Age | Commit message (Expand) | Author |
---|---|---|
2023-01-10 | Extend cl image support to input and output tensors | Gian Marco Iodice |
2022-10-07 | Workaround CL compiler issue on FP16 | Viet-Hoa Do |
2022-10-06 | Rework DepthwiseConvolution heuristic on OpenCL | Gian Marco Iodice |
2022-09-07 | Optimize depthwise convolution on OpenCL | Gian Marco Iodice |
2022-06-27 | Implement new Elementwise Dynamic Fusion Operators: Div, Floor | Michalis Spyrou |
2021-11-29 | Use loop unrolling only when the kernel height is less than 5 | Gian Marco Iodice |
2021-11-17 | Improve start-up time for depthwise convolution | Sheri Zhang |
2021-10-14 | Implement CLDirectConv3D f32/f16 | Giorgio Arena |
2021-07-25 | Reorganize the kernels into nhwc, nchw and common folders | Adnan AlSinan |