[Upstream fix] Fix after #23041 from upstream (#87)

adobrzyn · web-flow · commit 0492c5598089 · 2025-08-19T14:17:00.000+02:00
Culprit PR: vllm-project/vllm#23041 Signed-off-by: Agata Dobrzyniewicz <adobrzyniewicz@habana.ai>
diff --git a/vllm_gaudi/v1/worker/hpu_model_runner.py b/vllm_gaudi/v1/worker/hpu_model_runner.py
@@ -1906,7 +1906,6 @@ def execute_model(
             req_id_to_index=self.input_batch.req_id_to_index,
             sampled_token_ids=postprocessed_sampled_token_ids,
             logprobs=logprobs,
-            spec_token_ids=None,
             prompt_logprobs_dict=prompt_logprobs_dict,  # type: ignore[arg-type]
             pooler_output=[],
         )

Original file line number	Diff line number	Diff line change
`@@ -1906,7 +1906,6 @@ def execute_model(`
`1906`	`1906`	`req_id_to_index=self.input_batch.req_id_to_index,`
`1907`	`1907`	`sampled_token_ids=postprocessed_sampled_token_ids,`
`1908`	`1908`	`logprobs=logprobs,`
`1909`		`- spec_token_ids=None,`
`1910`	`1909`	`prompt_logprobs_dict=prompt_logprobs_dict, # type: ignore[arg-type]`
`1911`	`1910`	`pooler_output=[],`
`1912`	`1911`	`)`