a tiny bug fix missing default_training_args

#319

hello , there is a tiny bug in function "get_default_train_args" when num_layers != 6. A default_training_args which is a empy dict is missing.

icefire080 changed pull request status to closed

Sign up or log in to comment