GitLab

fix _forward_encoder_attention
remove medusa
set VLLM_PCIE_USE_CUSTOM_ALLREDUCE=1

[gpt-oss-120b] MoE configs for BW1000 TP=8

实现了用于优化张量计算的 rms_mrope_fuse 和 rms_mrope_fuse_fake 方法
更新了 forward：在满足条件时走新的 M-RoPE 融合路径
增强了 Qwen3MoeModel 对动态参数维度的支持，以适配该功能

update weights_not_loaded and flash_mla_with_kvcache
update paged_mqa_logits

Signed-off-by: sarathc-cerebras <sarath.chandran@cerebras.net>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
(cherry picked from commit 28d15ab5)

(cherry picked from commit 74a1ac38)

(cherry picked from commit e3fc374a)

Signed-off-by: Yan Ma <yan.ma@intel.com>
(cherry picked from commit 4f735bab)

Signed-off-by: jiang1.li <jiang1.li@intel.com>
Signed-off-by: Li, Jiang <bigpyj64@gmail.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>
(cherry picked from commit 0cd53536)

修复CompressedTensorsLinearMethod中的w4a16的冲突问题
feat(moe): add Marlin W16A16 fused MoE behind VLLM_USE_MARLIN_W16A16_MOE
replace the fp8_mqa_logits and fp8_paged_mqa_logits interfaces in deepgemm with mqa_logits and paged_mqa_logits from lightop

Signed-off-by: Roger Wang <hey@rogerw.io>
Co-authored-by: Sun Kim <sunytokki@gmail.com>
(cherry picked from commit f5f51e59)

Signed-off-by: Isotr0py <mozf@mail2.sysu.edu.cn>
(cherry picked from commit e94384bb)

Signed-off-by: shen-shanshan <467638484@qq.com>
Co-authored-by: gcanlin <canlinguosdu@gmail.com>
Co-authored-by: TJian <tunjian.tan@embeddedllm.com>
(cherry picked from commit 3bd9c491)

Signed-off-by: Matthew Bonanni <mbonanni@redhat.com>
Co-authored-by: gemini-code-assist[bot] <176961590+gemini-code-assist[bot]@users.noreply.github.com>

Signed-off-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>

Signed-off-by: Max Hu <hyoung2991@gmail.com>