minke.yu/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 12 Packages Projects Releases Wiki Activity

Add DeepSeek3.2 and GlmMoeDsa into moe tune (#18876)

Browse Source 
Co-authored-by: luoyuan.luo <luoyuan.luo@antgroup.com>
This commit is contained in:
Yuan Luo
2026-03-10 17:12:58 +08:00
committed by GitHub
co-authored by luoyuan.luo
parent ac07a6d439
commit 751c454099
1 changed files with 4 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
benchmark/kernels/fused_moe_triton/common_utils.py
+4
View File
@@ -86,7 +86,9 @@ def get_model_config(
elif architecture in [ elif architecture in [
"DeepseekV2ForCausalLM", "DeepseekV2ForCausalLM",
"DeepseekV3ForCausalLM", "DeepseekV3ForCausalLM",
"DeepseekV32ForCausalLM",
"Glm4MoeForCausalLM", "Glm4MoeForCausalLM",
"GlmMoeDsaForCausalLM",
"MistralLarge3ForCausalLM", "MistralLarge3ForCausalLM",
]: ]:
E = (config.n_routed_experts // ep_size) + ( E = (config.n_routed_experts // ep_size) + (
@@ -95,7 +97,9 @@ def get_model_config(
or architecture or architecture
not in [ not in [
"DeepseekV3ForCausalLM", "DeepseekV3ForCausalLM",
"DeepseekV32ForCausalLM",
"Glm4MoeForCausalLM", "Glm4MoeForCausalLM",
"GlmMoeDsaForCausalLM",
"MistralLarge3ForCausalLM", "MistralLarge3ForCausalLM",
] ]
else 1 else 1