GitLab

Co-authored-by: mgoin <mgoin64@gmail.com>

Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>

Signed-off-by: mgoin <mgoin64@gmail.com>

Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>

Signed-off-by: Caleb_Du <Caleb_Du@zju.edu.cn>

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>

Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
Co-authored-by: mgoin <mgoin64@gmail.com>

[FEAT] [ROCm] [Embedding] Add encoder-only model support into ROCm Flash Attention to enable embedding models. (#14664)
Signed-off-by: tjtanaa <tunjian.tan@embeddedllm.com>

Signed-off-by: Sage Moore <sage@neuralmagic.com>

Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>
Co-authored-by: mgoin <mgoin64@gmail.com>

Signed-off-by: mgoin <mgoin64@gmail.com>

sgl_moe_align_block_size is based on:


https://github.com/sgl-project/sglang/commit/ded9fcd09a43d5e7d5bb31a2bc3e9fc21bf65d2a

moe_align_block_size is based on:


https://github.com/sgl-project/sglang/commit/ba5112ff691d791a9e38c6c71f59324a5fcb49d0

Signed-off-by: Yang Chen <yangche@fb.com>

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>

Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>

[Kernel] optimize moe_align_block_size for cuda graph and large num_experts (e.g. DeepSeek-V3) (#12222)
Signed-off-by: Jinzhen Lin <linjinzhen@hotmail.com>
Co-authored-by: Michael Goin <mgoin@redhat.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>

Signed-off-by: mgoin <michael@neuralmagic.com>
Co-authored-by: mgoin <michael@neuralmagic.com>
Co-authored-by: robertgshaw2-neuralmagic <rshaw@neuralmagic.com>

Signed-off-by: charlifu <charlifu@amd.com>

Co-authored-by: Dipika <dipikasikka1@gmail.com>
Co-authored-by: Dipika Sikka <ds3822@columbia.edu>

Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>

Co-authored-by: mgoin <michael@neuralmagic.com>

Co-authored-by: lwilkinson@neuralmagic.com

Co-authored-by: Dipika <dipikasikka1@gmail.com>

Co-authored-by: ElizaWszola <eliza@neuralmagic.com>

Co-authored-by: ElizaWszola <eliza@neuralmagic.com>

This PR enables the fused topk_softmax kernel used in moe layer for HIP