LightEval is a lightweight LLM evaluation suite that Hugging Face has been using internally with the recently released LLM data processing library datatrove and LLM training library nanotron.
MIT License
467
stars
54
forks
source link
Added Namespace parameter for InferenceEndpoints, added option for passing model config directly #147
1] I added the namespace parameter to the endpoint model config, to allow deploying endpoints from an organization that isn't your private user.
Added the namespace property to the class
Added a static method nullable_keys() so that the check for a complete config when creating the endpoint can ignore those.
2] I added the option to specify the model config directly from python code, with a simple check to see if it was given. (Note: I unindented everything after the yaml.safe_load, since after the yaml was loaded there was no need to remain inside the scope of accessing the file).
1] I added the
namespace
parameter to the endpoint model config, to allow deploying endpoints from an organization that isn't your private user.nullable_keys()
so that the check for a complete config when creating the endpoint can ignore those.2] I added the option to specify the model config directly from python code, with a simple check to see if it was given. (Note: I unindented everything after the
yaml.safe_load
, since after the yaml was loaded there was no need to remain inside the scope of accessing the file).Sending you sample code to test on slack.