fix: fix MLA for ShardedModelLoader/RemoteModelLoader (#6287)

Signed-off-by: wangyu <wangyu.steph@bytedance.com>
2025-08-29 07:10:09 +08:00
parent a38c149758
commit 9f81d741a2
8 changed files with 37 additions and 35 deletions
--- a/examples/runtime/engine/save_remote_state.py
+++ b/examples/runtime/engine/save_remote_state.py
@@ -14,8 +14,7 @@ python save_remote_state.py \
 Then, the model can be loaded with

 llm = Engine(
-    model_path="/path/to/save",
-    --remote-model-url [protocol]://[host]:[port]/[model_name],
+    model_path="[protocol]://[host]:[port]/[model_name]",
    tensor_parallel_size=8,
 )
 """