[Bugfix] Fix loading of fine-tuned models based on Phi-3-Small (#12689)

Signed-off-by: Michael Greenbaum <mgreenbaum@microsoft.com>
Co-authored-by: Michael Greenbaum <mgreenbaum@microsoft.com>

vllm/model_executor/models/phi3_small.py

@@ -476,6 +476,8 @@ class Phi3SmallForCausalLM(nn.Module, SupportsPP):
                 continue
             if is_pp_missing_parameter(name, self):
                 continue
+            if "lm_head.weight" in name and self.config.tie_word_embeddings:
+                continue
             param = params_dict[name]
             weight_loader = getattr(param, "weight_loader",
                                     default_weight_loader)