Modularize fused experts and integrate PPLX kernels (#15956)

2025-05-14 16:11:54 -04:00
parent 418d2f8bfb
commit f9c069c85e
42 changed files with 3830 additions and 660 deletions
--- a/vllm/v1/worker/gpu_worker.py
+++ b/vllm/v1/worker/gpu_worker.py
@@ -341,7 +341,8 @@ def init_worker_distributed_environment(
                                 distributed_init_method, local_rank)

    ensure_model_parallel_initialized(parallel_config.tensor_parallel_size,
-                                      parallel_config.pipeline_parallel_size)
+                                      parallel_config.pipeline_parallel_size,
+                                      parallel_config.enable_expert_parallel)

    ensure_kv_transfer_initialized(vllm_config)