wangyxbh 5027739f2c [CPU] Use covt_e4m3_bf16 to optim BF16 to FP8 convert (#12191)

2025-11-14 17:36:51 +08:00

..

allreduce

[bug][rocm]fix qr when variable inp (#11609)

2025-11-11 01:43:48 -08:00

attention

…

cpu

[CPU] Use covt_e4m3_bf16 to optim BF16 to FP8 convert (#12191)

2025-11-14 17:36:51 +08:00

cutlass_extensions

…

elementwise

[sgl-kernel][Deepseek V3.2] Add row_starts to topk kernel (#12582)

2025-11-07 18:33:27 -08:00

expert_specialization

…

gemm

Support hidden_dim % 4 == 0 in per_token_quant_fp8 (#12883)

2025-11-10 17:13:14 +08:00

grammar

…

kvcacheio

…

mamba

Update GDN causal conv1d cuda kernel - prepare for new changes (#13188)

2025-11-13 14:09:47 -08:00

memory

Migrate weak_ref_tensor to sgl-kernel (#12505)

2025-11-02 10:55:39 +08:00

moe

…

quantization/gguf

…

spatial

…

speculative

move all get_stream in sgl_kernel to c++ to reduce the launch overhead (#12521)

2025-11-02 13:15:05 -08:00

common_extension_rocm.cc

Fix errors of page head kernels in sgl-kernel for ROCm (#12604)

2025-11-10 15:15:50 +08:00

common_extension.cc

[sgl-kernel][Deepseek V3.2] Add row_starts to topk kernel (#12582)

2025-11-07 18:33:27 -08:00

flash_extension.cc

…

flashmla_extension.cc

[sgl-kernel] support custom fp8 flashmla kernel (#13087)

2025-11-13 12:45:21 -08:00

spatial_extension.cc

…