Merge remote-tracking branch 'origin/v0.5.4_dev' into sglang_v0.5.5

a1175a4e · maxiao1 · 0c006b88 · 31653dd9 · a1175a4e · a1175a4e
Commit a1175a4e authored Nov 22, 2025 by maxiao1
20 changed files
--- a/python/sglang/srt/model_executor/model_runner.py
+++ b/python/sglang/srt/model_executor/model_runner.py
--- a/python/sglang/srt/models/deepseek_v2.py
+++ b/python/sglang/srt/models/deepseek_v2.py
--- a/python/sglang/srt/models/qwen3_next.py
+++ b/python/sglang/srt/models/qwen3_next.py
--- a/python/sglang/srt/profile/prof.py
+++ b/python/sglang/srt/profile/prof.py
--- a/python/sglang/srt/server_args.py
+++ b/python/sglang/srt/server_args.py
--- a/python/sglang/srt/speculative/draft_utils.py
+++ b/python/sglang/srt/speculative/draft_utils.py
--- a/python/sglang/srt/speculative/eagle_info.py
+++ b/python/sglang/srt/speculative/eagle_info.py
--- a/python/sglang/srt/speculative/eagle_info_v2.py
+++ b/python/sglang/srt/speculative/eagle_info_v2.py
--- a/sgl-kernel/csrc/allreduce/custom_all_reduce_hip.cuh
+++ b/sgl-kernel/csrc/allreduce/custom_all_reduce_hip.cuh
--- a/sgl-kernel/csrc/attention/merge_attn_states.cu
+++ b/sgl-kernel/csrc/attention/merge_attn_states.cu
--- a/sgl-kernel/csrc/common_extension_rocm.cc
+++ b/sgl-kernel/csrc/common_extension_rocm.cc
--- a/sgl-kernel/csrc/gemm/qserve_w4a8_per_chn_gemm.cu
+++ b/sgl-kernel/csrc/gemm/qserve_w4a8_per_chn_gemm.cu
--- a/sgl-kernel/csrc/gemm/qserve_w4a8_per_group_gemm.cu
+++ b/sgl-kernel/csrc/gemm/qserve_w4a8_per_group_gemm.cu
--- a/sgl-kernel/csrc/kvcacheio/transfer.cu
+++ b/sgl-kernel/csrc/kvcacheio/transfer.cu
--- a/sgl-kernel/csrc/moe/moe_align_kernel.cu
+++ b/sgl-kernel/csrc/moe/moe_align_kernel.cu
--- a/sgl-kernel/csrc/moe/moe_topk_softmax_kernels.cu
+++ b/sgl-kernel/csrc/moe/moe_topk_softmax_kernels.cu
--- a/sgl-kernel/csrc/quantization/gguf/ggml-common.h
+++ b/sgl-kernel/csrc/quantization/gguf/ggml-common.h
--- a/sgl-kernel/include/sgl_kernel_ops.h
+++ b/sgl-kernel/include/sgl_kernel_ops.h
--- a/sgl-kernel/include/utils.h
+++ b/sgl-kernel/include/utils.h
--- a/sgl-kernel/python/sgl_kernel/flash_mla.py
+++ b/sgl-kernel/python/sgl_kernel/flash_mla.py