Hi.
I wrote export script to Core ML referring export_model_pytorch.py.
Now I wonder if I should use -use-swa option or not.
I tried both and compared a few values in the weights. they seems same.
Is b18c384nbt.bin.gz generated using -use-swa option?
Thanks.
Hi.
I wrote export script to Core ML referring export_model_pytorch.py.
Now I wonder if I should use -use-swa option or not.
I tried both and compared a few values in the weights. they seems same.
Is b18c384nbt.bin.gz generated using -use-swa option?
Thanks.