Class GPULlama3ResponseParser.StreamingParser
java.lang.Object
dev.langchain4j.model.gpullama3.GPULlama3ResponseParser.StreamingParser
- Enclosing class:
GPULlama3ResponseParser
Parser for handling streaming responses with real-time thinking content separation.
This parser detects thinking content as tokens are generated and routes it to the appropriate handler methods (onPartialThinking vs onPartialResponse). The thinking tags are preserved and streamed as part of the thinking content.
-
Constructor Summary
ConstructorsConstructorDescriptionStreamingParser
(StreamingChatResponseHandler handler, org.beehive.gpullama3.model.Model model) Creates a new streaming parser. -
Method Summary
Modifier and TypeMethodDescriptionvoid
onToken
(int tokenId) Processes each token as it's generated by the model.
-
Constructor Details
-
StreamingParser
public StreamingParser(StreamingChatResponseHandler handler, org.beehive.gpullama3.model.Model model) Creates a new streaming parser.- Parameters:
handler
- the streaming response handlermodel
- the GPULlama3 model instance for token decoding
-
-
Method Details
-
onToken
public void onToken(int tokenId) Processes each token as it's generated by the model.- Parameters:
tokenId
- the token ID generated by the model
-