Hello, I am new to pytorch-lightning and I have encountered a problem with the stable-diffusion code from MLPerf which seems to use an old version of lightning. So, the problem is, the add_argparse_args
is no longer available. I wonder how to migrate this piece of code to the new version.
def get_parser(**parser_kwargs):
parser = argparse.ArgumentParser(**parser_kwargs)
parser.add_argument(
"-n",
"--name",
type=str,
const=True,
default="",
nargs="?",
help="postfix for logdir",
)
return parser
if __name__ == "__main__":
parser = get_parser()
parser = Trainer.add_argparse_args(parser)