fix(moe): relocate orphan tuned configs after #23019 (#24004)

Browse Source

...

This commit is contained in:

hhwxw

2026-04-29 02:00:13 -07:00

committed by GitHub

parent 6c7b242181

commit d9270b8c6a

35 changed files with 5 additions and 5 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download Patch File Download Diff File Expand all files Collapse all files

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_B200,dtype=fp8_w8a8.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_B200,dtype=fp8_w8a8.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20,dtype=fp8_w8a8.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20,dtype=fp8_w8a8.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20,dtype=fp8_w8a8_down.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20,dtype=fp8_w8a8_down.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e,dtype=fp8_w8a8.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e,dtype=fp8_w8a8.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e,dtype=fp8_w8a8_down.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e,dtype=fp8_w8a8_down.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e_down.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20-3e_down.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20_down.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=192,N=192,device_name=NVIDIA_H20_down.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=1792,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=1792,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=1792,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=1792,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=224,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=224,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=224,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=224,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=448,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=448,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=448,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=448,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=896,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=896,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=32,N=896,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=32,N=896,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=1536,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=1536,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=1536,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=1536,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=192,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=192,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=192,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=192,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=384,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=384,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=384,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=384,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=768,device_name=NVIDIA_B200.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=768,device_name=NVIDIA_B200.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_5_1/E=64,N=768,device_name=NVIDIA_H100_80GB_HBM3.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_5_1/E=64,N=768,device_name=NVIDIA_H100_80GB_HBM3.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=32,N=1792,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=32,N=1792,device_name=AMD_Instinct_MI325X.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=32,N=224,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=32,N=224,device_name=AMD_Instinct_MI325X.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=32,N=448,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=32,N=448,device_name=AMD_Instinct_MI325X.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=32,N=896,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=32,N=896,device_name=AMD_Instinct_MI325X.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=64,N=1536,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=64,N=1536,device_name=AMD_Instinct_MI325X.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=64,N=192,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=64,N=192,device_name=AMD_Instinct_MI325X.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=64,N=384,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=64,N=384,device_name=AMD_Instinct_MI325X.json

python/sglang/srt/layers/moe/fused_moe_triton/configs/triton_3_6_0/E=64,N=768,device_name=AMD_Instinct_MI325X.json → python/sglang/srt/layers/moe/moe_runner/triton_utils/configs/triton_3_6_0/E=64,N=768,device_name=AMD_Instinct_MI325X.json