Closed bilgehanertan closed 6 months ago
Hi! Thanks for this PR!
You also need to add it to the InferenceEndpointModelConfig
class, and grab it from the kwargs in the main to initialize the class :)
I am not sure what you mean about the kwargs part. Isn't it directly load model_dtype?
You need to make sure that the argument is passed at the different places where the model config is created - I can give you more pointers tmr if you need :)
Hi, A more in depth version of this PR was done in #124 (notably by supporting all possible dtypes), so we will probably merge the other one. Thanks a lot for taking a look :)
Resolves #117