This kind of work has much more of a problem with finite training data than language models do - there are about 150,000 known protein structures, though a lot of them are very close homologues, often ...
一些您可能无法访问的结果已被隐去。
显示无法访问的结果一些您可能无法访问的结果已被隐去。
显示无法访问的结果