We have to define stride explicitly in torch avg_pool function or deal with the default None type of stride in tvm frontend pytorch.
確定! 回上一頁