Chen Qian
Chen Qian
I cannot reproduce the issue: ``` import dspy dspy.settings.configure(lm=dspy.LM("openai/gpt-4o-mini")) math = dspy.ChainOfThought(signature="question -> answer: float") print(math(question="Two dice are tossed. What is the probability that the sum equals two?")) ``` produces...
@Krishn1412 assigned to you!
@buchannan162 Sometimes model doesn't follow the instruction well. Try adding this to your signature's instruction (docstring): ``` Make sure when you decide the tool arguments, the section header should use...
@simonfrey Thanks for reporting the issue! We are actually working on supporting the pydantic constraints in DSPy.
Yes we support this now but in a soft way: https://github.com/stanfordnlp/dspy/blob/6ee8cdca4bf3283a8bde6b92c628cbbd0851fe9b/dspy/signatures/field.py#L14, which means for a small model it's still possible for it not following this rule, but for large models...
I am not sure if we need this, is it a valid/common use case to generate passing the end token?
Yea this should not block our release, I will write something later.
@arnavsinghvi11 Ah I will delete this PR later, this PR is written for some demo purposes.
@xihajun Junfan, sorry for the late reply, I have been in a global trip since Feb 15th, so missed lots of threads. To your question - We are still adjusting...
@Cyber-Machine Thanks for your interest! A guide on generation algos sounds good! Do you have a skeleton in your mind, like what algorithms do you want to cover? For `keras_nlp.models.GPT2CausalLM`,...