解决RuntimeError: index_copy_(): self and source expected to have the same dtype, but got (self) Int a

在运行如下代码时遇到报错

it.index_copy_(0, sample_ind, torch.multinomial(prob_prev, 1).view(-1).index_select(0, sample_ind))

RuntimeError: index_copy(): self and source expected to have the same dtype, but got (self) Int and (source) Long

原因是因为数据类型不对，一个时int，一个是long，修改一下数据类型就行。将代码修改为如下代码即可运行。

it.index_copy_(0, sample_ind, torch.multinomial(prob_prev, 1).view(-1).index_select(0, sample_ind).int())