Error converting kykim/gpt3-kor-small_based_on_gpt2

#51
by jonomon - opened

Conversion Settings:

        Model: kykim/gpt3-kor-small_based_on_gpt2
        Task: text-generation
        Compute Units: ComputeUnit.ALL
        Precision: float16
        Tolerance: None
        Push to: None

        Error: Output values do not match between reference model and Core ML exported model: Got max absolute difference of: nan
        

The conversion works with Precision = float32, but fails with float16.

Thank you

Sign up or log in to comment