[Model Runner V2] Fix warmup for pipeline parallel (#36280)

Signed-off-by: Nick Hill <nickhill123@gmail.com>

vllm/v1/worker/gpu_worker.py

@@ -584,7 +584,7 @@ class Worker(WorkerBase):
 
         if self.use_v2_model_runner:
             # V2: Run full execute_model + sample_tokens to JIT compile triton kernels.
-            warmup_kernels(self.model_runner)
+            warmup_kernels(self.model_runner, self.execute_model, self.sample_tokens)
         elif get_pp_group().is_last_rank:
             # V1: Warm up sampler and preallocate memory buffer for logits and other
             # sampling related tensors of max possible shape to avoid memory