From 719e34cb9113aacb4747e972097fc72135591cbf Mon Sep 17 00:00:00 2001
From: Roger Wang <ywang@roblox.com>
Date: Fri, 15 Nov 2024 19:51:19 -0800
Subject: [PATCH] fix xverse

Signed-off-by: Roger Wang <ywang@roblox.com>
---
 vllm/model_executor/models/xverse.py | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/vllm/model_executor/models/xverse.py b/vllm/model_executor/models/xverse.py
index 949c2e99a91b0..51172d8782a70 100644
--- a/vllm/model_executor/models/xverse.py
+++ b/vllm/model_executor/models/xverse.py
@@ -352,9 +352,11 @@ def forward(
         kv_caches: List[torch.Tensor],
         attn_metadata: AttentionMetadata,
         intermediate_tensors: Optional[IntermediateTensors] = None,
+        inputs_embeds: Optional[torch.Tensor] = None,
     ) -> Union[torch.Tensor, IntermediateTensors]:
         hidden_states = self.model(input_ids, positions, kv_caches,
-                                   attn_metadata, intermediate_tensors)
+                                   attn_metadata, intermediate_tensors,
+                                   inputs_embeds)
         return hidden_states
 
     def compute_logits(