Target dataset is large and similar to the original training dataset