ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-15 02:27:57 +00:00
Code Issues Packages Projects Releases Wiki Activity
Files
0ac91713ae8a84dbbb73d848ab526b33d39494aa
composable_kernel/include/ck/tensor_operation/gpu
History
huaiguxu 0ac91713ae Huaiguxu/moe fp8 pertoken scale fix (#2391) 
* fix pertoken_scale a_scale dimension

* clang-format

* Fix moe_gemm2_fp8 perTokenScale reference and example.

[ROCm/composable_kernel commit: e1c5172fdb]
2025-06-27 10:24:34 +08:00
..
block
fix moe i4 bug from aiter (#2339)
2025-06-24 14:51:29 +08:00
device
[CK][CONV] Support NCHW in class DeviceGroupedConvFwdMultipleABD_Xdl_CShuffle (#2375)
2025-06-26 08:32:39 +08:00
element
Grouped convolution forward with clamp (#2334)
2025-06-16 15:36:53 +02:00
grid
Huaiguxu/moe fp8 pertoken scale fix (#2391)
2025-06-27 10:24:34 +08:00
thread
Add MoE & FP8 Blockscale WP Kernels for GFX950 (#2297)
2025-06-12 09:25:59 +08:00
warp
Add MoE & FP8 Blockscale WP Kernels for GFX950 (#2297)
2025-06-12 09:25:59 +08:00