LangChain4j comments

Results 849 comments of


                                            LangChain4j

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

@jacksonhu which LLM are you using? As I've commented, this seems to be a misuse of tools (`@Tool` in LC4j), so I am not sure it is a good idea...

@daixi98 please note that many of the endpoints were implemented by @stephanj [here](https://github.com/stephanj/langchain4j-cohere/blob/master/src/main/java/dev/langchain4j/model/cohere/CohereApi.java) and will be migrated to the main repo soon. This PR is about [chat enadpoint](https://docs.cohere.com/reference/chat).

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

@juinjonn, a few questions: 1. Could you please provide a few examples (questions and tools) of how this will be used? 2. How do you expect to get the tool...

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

Hi @juinjonn, thank you for detailed explanation! 1. This makes sense! I am wondering if for this use case if would be beneficial to create a new class or maybe...

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

> I think your solution is excellent and can satisfy this case. However, there is one issue: in such scenarios, the focus is more on the tool’s results. If the...

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

> It seems more reasonable to configure this on the @tool annotation. This way, for a single AiService, its tools can let the user control which tools need to be...

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

@jacksonhu I am not sure I fully understand. Let's imagine LLM decides to call 2 tools, result of the first one should be returned directly, result of the second one...

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

@juinjonn please feel free to open the PR! Or perharps @jacksonhu could adjust this PR? Do we agree on the following solution? - Add `returnDirectly` attribute to the `@Tool` annotation...

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

Hi @juinjonn, do you still plan to submit a PR? 🙏

LangChain4j

[FEATURE] Cohere R/R+

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

[FEATURE] Cohere R/R+

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count

add attribute “returnAsFinalAnswer” to Tool annotation ,let llm can use tool funcion return value as the output of the model directly,to reduce llm request count