[TPU] Avoid Triton Import (#15589)

Signed-off-by: rshaw@neuralmagic.com <robertgshaw2@gmail.com>

Author: robertgshaw2-redhat

vllm/model_executor/layers/fused_moe/layer.py

@@ -16,8 +16,6 @@
 from vllm.forward_context import ForwardContext, get_forward_context
 from vllm.logger import init_logger
 from vllm.model_executor.custom_op import CustomOp
-from vllm.model_executor.layers.fused_moe.rocm_aiter_fused_moe import (
-    is_rocm_aiter_moe_enabled, shuffle_weights)
 from vllm.model_executor.layers.quantization.base_config import (
     QuantizationConfig, QuantizeMethodBase)
 from vllm.model_executor.utils import set_weight_attrs
@@ -119,7 +117,9 @@ def process_weights_after_loading(self, layer: torch.nn.Module) -> None:
         layer.w2_weight = torch.nn.Parameter(self._maybe_pad_weight(
             layer.w2_weight.data),
                                              requires_grad=False)
-
+        # Lazy import to avoid importing triton.
+        from vllm.model_executor.layers.fused_moe.rocm_aiter_fused_moe import (
+            is_rocm_aiter_moe_enabled, shuffle_weights)
         if is_rocm_aiter_moe_enabled():
             # reshaping weights is required for aiter moe kernel.
             shuffled_w13, shuffled_w2 = shuffle_weights(

vllm/model_executor/layers/quantization/fp8.py

@@ -13,9 +13,6 @@
 from vllm.logger import init_logger
 from vllm.model_executor.layers.fused_moe import (FusedMoE, FusedMoEMethodBase,
                                                   FusedMoeWeightScaleSupported)
-from vllm.model_executor.layers.fused_moe.rocm_aiter_fused_moe import (
-    expand_weights, is_rocm_aiter_block_scaled_moe_enabled,
-    is_rocm_aiter_moe_enabled, shuffle_weights)
 from vllm.model_executor.layers.linear import (LinearBase, LinearMethodBase,
                                                UnquantizedLinearMethod)
 from vllm.model_executor.layers.quantization.base_config import (
@@ -532,6 +529,11 @@ def create_weights(self, layer: Module, num_experts: int, hidden_size: int,
             layer.w2_input_scale = None
 
     def process_weights_after_loading(self, layer: Module) -> None:
+        # Lazy import to avoid importing triton too early.
+        from vllm.model_executor.layers.fused_moe.rocm_aiter_fused_moe import (
+            expand_weights, is_rocm_aiter_block_scaled_moe_enabled,
+            is_rocm_aiter_moe_enabled, shuffle_weights)
+
         # TODO (rob): refactor block quant into separate class.
         if self.block_quant:
             assert self.quant_config.activation_scheme == "dynamic"