Open
Description
Thanks for your excellent work. However, I found some weird performance of the tool when working with the in-house model client.
Firstly, I define the client:
from openai import OpenAI
client = OpenAI(base_url="IN-HOUSE IP address", api_key="IN-HOUSE KEY")
And I register the model by using the config:
ell.config.register_model("CUSTOM", client)
Then I use the ell.simple:
ell.init(store="./ell_logdir", verbose=True)
@ell.simple(model="CUSTOM", client=client)
def my_lmp(prompt: str):
"""You are a helpful assistant."""
return f"Respond to: {prompt}"
Finally, I use this simple function:
temp = my_lmp("Sam")
logger.info(temp)
logger.info("Finish")
But I got nothing:
By the way, if I do not use the ell to help me record the prompt, I can get the correct response.
Metadata
Metadata
Assignees
Labels
No labels