Jason B. Koh comments

Results 38 comments of


                                            Jason B. Koh

Expose request metadata in the generated ```ServiceBase```

Are there any other suggestions on this issue? I forked betterproto and did something similar to @iamnoah's solution, but if there is a better practice suggested by the community, I'd...

Compact serialization

I'd be interested in this issue too. I understand it's easier to assign ids to blank nodes for serialization, but one of the upsides of JSON-LD is the nested format...

Just want to chime in to support this thread. I just want to simply add JWT as a security scheme, which is not defined in [Swagge2.0](https://swagger.io/docs/specification/2-0/authentication/) but in [OpenAPI3.0](https://github.com/OAI/OpenAPI-Specification/blob/master/versions/3.0.0.md#jwt-bearer-sample).

Expose request metadata in the generated ```ServiceBase```

> I guess I should switch back to grpclib @parhawm This solution works great: https://github.com/danielgtaylor/python-betterproto/pull/234#issuecomment-1089905019

Fix: OpenAIFunctionsAgent often returns python as a tool in planning, when argument parsing fails

I didn't fix `OpenAIMultiFunctionAgent` just yet. If all agrees with this approach, I will update the other part too.

Fix: OpenAIFunctionsAgent often returns python as a tool in planning, when argument parsing fails

I thought `python` was some default tool, but it turned out GPT is throwing out a random tool and I got this result from the agent ``` Function: python is...

Fix: OpenAIFunctionsAgent often returns python as a tool in planning, when argument parsing fails

So, is this just OpenAI's transient behaviors trying to use undefined tools? Shouldn't we capture the errors by checking the tool names instead of raising a unexpected parsing error? And...

Fix: OpenAIFunctionsAgent often returns python as a tool in planning, when argument parsing fails

not relevant anymore

Multi core support for training on large number of instances

I wonder if this development of multicore CRF has been dead or not. I am dying for such feature.

Multi core support for training on large number of instances

@usptact @tianjianjiang Thanks for the information! I wish I could have exploited the cores with PyCRFSuite, but I can switch to the pointer. Thank you all.