biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

Refactor pplx init logic to make it modular (prepare for deepep) (#18200)

Browse Source 
Signed-off-by: youkaichao <youkaichao@gmail.com>
This commit is contained in:
youkaichao
2025-05-23 23:43:43 +08:00
committed by GitHub
parent 022d8abe29
commit 6a7988c55b
16 changed files with 300 additions and 287 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
vllm/v1/worker/gpu_worker.py
View File

@@ -348,8 +348,7 @@ 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.enable_expert_parallel)
parallel_config.pipeline_parallel_size)
ensure_kv_transfer_initialized(vllm_config)

3
vllm/v1/worker/tpu_worker.py
View File

@@ -265,8 +265,7 @@ def init_tpu_worker_distributed_environment(
backend="gloo",
)
ensure_model_parallel_initialized(parallel_config.tensor_parallel_size,
parallel_config.pipeline_parallel_size,
parallel_config.enable_expert_parallel)
parallel_config.pipeline_parallel_size)
try: