Clean up some Qwen3-Next and deterministic code (#11585)

python/sglang/srt/configs/mamba_utils.py

@@ -70,7 +70,7 @@ class Mamba2StateShape:
 
         # These are not TP-ed as they depend on A, dt_bias, D
         # - they are typically small
-        #   e.g., (h_heads, head_dim, state_size) = (128, 64, 128)
+        #   e.g., QWen3-Next: (32, 128, 128)
         temporal_state_shape = (divide(num_heads, tp_world_size), head_dim, state_size)
         return Mamba2StateShape(
             conv=conv_state_shape,

python/sglang/srt/configs/qwen3_next.py

@@ -27,12 +27,9 @@ from sglang.srt.layers.dp_attention import get_attention_tp_size
 logger = logging.get_logger(__name__)
 
 
-# NOTE: HybridLayerType
 class HybridLayerType(enum.Enum):
     full_attention = "attention"
-    swa_attention = "swa_attention"
     linear_attention = "linear_attention"
-    mamba2 = "mamba"
 
 
 class Qwen3NextConfig(PretrainedConfig):

python/sglang/srt/models/falcon_h1.py

@@ -450,13 +450,6 @@ class FalconH1Model(nn.Module):
         return hidden_states
 
 
-class HybridLayerType(enum.Enum):
-    full_attention = "attention"
-    swa_attention = "swa_attention"
-    linear_attention = "linear_attention"
-    mamba2 = "mamba"
-
-
 class FalconH1ForCausalLM(nn.Module):
     fall_back_to_pt_during_load = False
 

python/sglang/test/test_deterministic.py

@@ -226,10 +226,6 @@ def send_prefix(args, batch_size: int, prompts: List[str]):
 
 
 def test_deterministic(args):
-    # First do some warmups
-    for i in range(3):
-        send_single(args, 16, args.profile)
-
     if args.test_mode == "single":
         # In single mode, we test the deterministic behavior by sending the same prompt in batch sizes ranging from 1 to n_trials.
         texts = []