Open-source LLM from HuggingFace Agent Support for json output parsing #130
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
yyiilluu
reviewed
Sep 13, 2023
|
|
||
| return response | ||
|
|
||
| @staticmethod | ||
| def _fix_message(clean_text: str) -> UserMessage: |
There was a problem hiding this comment.
can you make this is a helper function inside of _attempt_fix_and_generate?
| @@ -55,30 +54,94 @@ def format_output(self) -> Dict[str, Any]: | |||
|
|
|||
|
|
|||
| class AgentOutputParser(BaseModel): | |||
There was a problem hiding this comment.
i think there is a way to simplify this logics inside of this function
first json load, if run into exception, fix and generate, format into message for load_json_output, decrease retry, call load_json_output again,
if no exception, just return
|
No worries at all! I will work on your comments, thank you so much |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Hi,
I added HuggingFace
text-generationmodels support, which was failing because of the use ofChatOpenAI(OpenAI API Key error occurs) whenConversationalAgentapplies json output parsing throughConvoJSONOutputParserthat usesload_json_output. To do that, instead of calling ChatOpenAI in theexceptpart of theload_json_output, I am passing whatever llm model is given, either ChatOpenAI or HuggingFaceTextGenerationModel, just passing the provided llm inside to make it more general.I also added
max_retryto set the amount of retry that recursive function being applied to get the correct output that JSON parseable._fix_messageneeds to be improved to get better responses from models each time it is called. I added it as TO DO as well as a comment there.After
max_retrynumber is reached and if we couldn't get the proper json parseable response, I raised ValueError to inform user that selected open-source model is not good enough or is not able to generate json parseable output.I checked this modified code with ChatOpenAI and it is working with no issue. But, there is a lot of things to do to improve and set the boundaries for the open-source model use. This may mean one needs to specify and allow some specific open-source models only. Also, some model optimizations can be applied.