Hi,
I noticed that generation.py does the following in several places:
with llm_params(llm, temperature=self.config.lowest_temperature):
However it seems that for temperature to be passed to bedrock it would need to be:
with llm_params(llm, model_kwargs={'temperature': self.config.lowest_temperature}):
I'm not sure of the cleanest way to make it compatible with both, but just wanted to alert you to this issue. Thanks!