Update TensorRT-LLM backend (#735)

kaiyux · web-flow · commit 689a553a8641 · 2025-04-16T08:11:26.000+08:00
diff --git a/inflight_batcher_llm/src/model_instance_state.cc b/inflight_batcher_llm/src/model_instance_state.cc
@@ -623,7 +623,7 @@ executor::ExecutorConfig ModelInstanceState::getExecutorConfigFromParams()
 
             mExecutorLookaheadDecodingConfig
                 = executor::LookaheadDecodingConfig{windowSize, ngramSize, verificationSetSize};
-            decodingConfig.setLookaheadDecoding(mExecutorLookaheadDecodingConfig.value());
+            decodingConfig.setLookaheadDecodingConfig(mExecutorLookaheadDecodingConfig.value());
         }
         catch (std::exception const& e)
         {
diff --git a/tensorrt_llm b/tensorrt_llm
@@ -1 +1 @@
-Subproject commit 5aeef6d4c72df864c5f4a20cdb899f278fff66f4
+Subproject commit 258ae9c58ce1a16b6a8e681be3d7695f731b1886
diff --git a/tools/version.txt b/tools/version.txt
@@ -1 +1 @@
-fd606a27785406dbde65311f28fc4941e159706d
+13c75b7dbd781eba6ac093f9acb2f37baa034561

Original file line number	Diff line number	Diff line change
`@@ -623,7 +623,7 @@ executor::ExecutorConfig ModelInstanceState::getExecutorConfigFromParams()`
`623`	`623`
`624`	`624`	`mExecutorLookaheadDecodingConfig`
`625`	`625`	`= executor::LookaheadDecodingConfig{windowSize, ngramSize, verificationSetSize};`
`626`		`- decodingConfig.setLookaheadDecoding(mExecutorLookaheadDecodingConfig.value());`
	`626`	`+ decodingConfig.setLookaheadDecodingConfig(mExecutorLookaheadDecodingConfig.value());`
`627`	`627`	`}`
`628`	`628`	`catch (std::exception const& e)`
`629`	`629`	`{`