[Misc] Collect model support info in a single process per model (#9233)

2024-10-11 19:08:11 +08:00
parent cbc2ef5529
commit e808156f30
4 changed files with 236 additions and 167 deletions
--- a/vllm/engine/arg_utils.py
+++ b/vllm/engine/arg_utils.py
@@ -183,6 +183,8 @@ class EngineArgs:
    def __post_init__(self):
        if self.tokenizer is None:
            self.tokenizer = self.model
+
+        # Setup plugins
        from vllm.plugins import load_general_plugins
        load_general_plugins()