从tf改写到torch,tf的placeholders,在pytorch中怎么处理?

从tf改写到torch,tf的placeholders,在pytorch中怎么处理?