add tunning files for QWEN-3-NEXT (#10794)

2025-09-24 03:46:30 +08:00
parent 23632d350c
commit 984730b732
3 changed files with 297 additions and 1 deletions
--- a/benchmark/kernels/fused_moe_triton/tuning_fused_moe_triton.py
+++ b/benchmark/kernels/fused_moe_triton/tuning_fused_moe_triton.py
@@ -411,7 +411,11 @@ def main(args: argparse.Namespace):
        topk = config.num_experts_per_tok
        intermediate_size = config.intermediate_size
        shard_intermediate_size = 2 * intermediate_size // args.tp_size
-    elif config.architectures[0] in ["Qwen2MoeForCausalLM", "Qwen3MoeForCausalLM"]:
+    elif config.architectures[0] in [
+        "Qwen2MoeForCausalLM",
+        "Qwen3MoeForCausalLM",
+        "Qwen3NextForCausalLM",
+    ]:
        E = config.num_experts
        topk = config.num_experts_per_tok
        intermediate_size = config.moe_intermediate_size