chenchenghao/sglang
1
0
Fork 0
Code Issues Pull Requests Actions 11 Packages Projects Releases Wiki Activity

Feat/nemotron nano v3 support (#12690)

Browse Source
This commit is contained in:
roikoren755
2025-11-21 13:53:05 -08:00
committed by GitHub
parent a24aefe562
commit 1b48e1b974
13 changed files with 776 additions and 68 deletions
Download Patch File Download Diff File Expand all files Collapse all files
benchmark/kernels/fused_moe_triton/common_utils.py
+1 -1
View File
@@ -104,7 +104,7 @@ def get_model_config(
E = config.num_experts // ep_size
topk = config.num_experts_per_tok
intermediate_size = config.moe_intermediate_size
elif architecture in ["Glm4MoeForCausalLM"]:
elif architecture in ["Glm4MoeForCausalLM", "NemotronHForCausalLM"]:
E = config.n_routed_experts // ep_size
topk = config.num_experts_per_tok
intermediate_size = config.moe_intermediate_size