Disabling cache not working

[lestan]

Tried to set the cache flag to False as per documentation to force requests to not use cache. Not getting the results I expect - i.e. getting the same results each time.

Documentation on the Migration guide:

gpt_4o_mini = dspy.LM('openai/gpt-4o-mini', temperature=0.9, max_tokens=3000, stop=None, cache=False)

My code uses Ollama:
lm = dspy.LM("ollama_chat/granite3-moe:1b", provider="ollama", api_base="http://localhost:11434/", cache=False, launch_kwargs={'seed', 42}) dspy.configure(lm=llm)

When I call it to generate a random name, I get the same results from run to run with the following code:

llm("generate a random human person name. no descriptions. no famous people. not fantasy or science fiction based.")

Am I missing something in the configuration?

Appreciate any help!

Les