Update README.md

README.md

@@ -116,6 +116,7 @@ vLLM aslo supports OpenAI-compatible serving. See the [documentation](https://do
       targets="Linear",
       scheme="W4A16",
       ignore=["lm_head"],
+      sequential_targets=["Phi3DecoderLayer"],
       dampening_frac=0.01,
   )