dspy

dspy copied to clipboard

GSM8k with assertions notebook

1

Shangyint

source is from https://github.com/stanfordnlp/dspy/pull/342/files - update had gotten stale. updates written by @DanielUH2019

csellis

Save of compiled program with pydantic types raises TypeError

3

compiled.save("./output/" + "compiled_v1.json") File "/Users/.../lib/python3.12/site-packages/dspy/primitives/module.py", line 79, in save f.write(ujson.dumps(self.dump_state(), indent=2)) I got TypeError: SomeType ....is not JSON serializable class SomeType(BaseModel): Field1: str = Field(..., description=“xxxx”) Field2: Optional[int] = Field(None,...

juicesharp

(Fixable?) 400 Error with vLLM API - extra input

2

Howdy. It seems when running a vLLM server and then attempting to interact with it via `HFClientVLLM`, I get an error message. Here is how to reproduce: ```bash # Computer...

ccruttjr

[Bug] Together AI completions bugs

1

Hi, I believe there are a few bugs in the together API. When hitting the chat/completions endpoint, the response json I get doesn't contain an "output" field, but the Together...

adam-simple

Update FAQ to mention `DSP_CACHEDIR` for local cache.

1

tom-doerr

The _call_model function in `aws_models.py` uses the Bedrock syntax of `invoke_model()` rather than `invoke_endpoint()` which is the syntax for Sagemaker. Although the code in `aws_providers.py` is correct, it gets overwritten...

vdemchenko3

[Prem AI](https://premai.io) is a generative AI platform that lets developers seamlessly integrate generative AI solutions (LLM generation, fine-tuning, RAGs, etc) in best way possible. This issue proposes to add it...

Anindyadeep

I want to build an agent to answer questions using data stored in a pandas data-frame (similar to langchain's data-frame Agent but customised to my needs). I tried different ways...

giresg

when we set a rm in dspy, and we call the dspy.Retrieve, the following lines will be executed https://github.com/stanfordnlp/dspy/blob/b05394bb20722444648a33b4447156bcef202fd6/dsp/primitives/search.py#L8-L24 so we can see in L12, the `passages` return by retrievers...

wxywb