Summary

主要包括以下三種途徑：

使用獨立的函數；

使用torch.type()函數；

使用type_as(tesnor)將張量轉換為給定類型的張量。

使用獨立函數

import torchtensor = torch.randn(3, 5)print(tensor)# torch.long() 將tensor投射為long類型long_tensor = tensor.long()print(long_tensor)# torch.half()將tensor投射為半精度浮點類型half_tensor = tensor.half()print(half_tensor)# torch.int()將該tensor投射為int類型int_tensor = tensor.int()print(int_tensor)# torch.double()將該tensor投射為double類型double_tensor = tensor.double()print(double_tensor)# torch.float()將該tensor投射為float類型float_tensor = tensor.float()print(float_tensor)# torch.char()將該tensor投射為char類型char_tensor = tensor.char()print(char_tensor)# torch.byte()將該tensor投射為byte類型byte_tensor = tensor.byte()print(byte_tensor)# torch.short()將該tensor投射為short類型short_tensor = tensor.short()print(short_tensor)

-0.5841 -1.6370 0.1353 0.6334 -3.0761-0.2628 0.1245 0.8626 0.4095 -0.3633 1.3605 0.5055 -2.0090 0.8933 -0.6267[torch.FloatTensor of size 3x5] 0 -1 0 0 -3 0 0 0 0 0 1 0 -2 0 0[torch.LongTensor of size 3x5]-0.5840 -1.6367 0.1353 0.6333 -3.0762-0.2627 0.1245 0.8628 0.4094 -0.3633 1.3604 0.5054 -2.0098 0.8936 -0.6265[torch.HalfTensor of size 3x5] 0 -1 0 0 -3 0 0 0 0 0 1 0 -2 0 0[torch.IntTensor of size 3x5]-0.5841 -1.6370 0.1353 0.6334 -3.0761-0.2628 0.1245 0.8626 0.4095 -0.3633 1.3605 0.5055 -2.0090 0.8933 -0.6267[torch.DoubleTensor of size 3x5]-0.5841 -1.6370 0.1353 0.6334 -3.0761-0.2628 0.1245 0.8626 0.4095 -0.3633 1.3605 0.5055 -2.0090 0.8933 -0.6267[torch.FloatTensor of size 3x5] 0 -1 0 0 -3 0 0 0 0 0 1 0 -2 0 0[torch.CharTensor of size 3x5] 0 255 0 0 253 0 0 0 0 0 1 0 254 0 0[torch.ByteTensor of size 3x5] 0 -1 0 0 -3 0 0 0 0 0 1 0 -2 0 0[torch.ShortTensor of size 3x5]

其中，torch.Tensor、torch.rand、torch.randn 均默認生成 torch.FloatTensor型 ：

import torchtensor = torch.Tensor(3, 5)assert isinstance(tensor, torch.FloatTensor)tensor = torch.rand(3, 5)assert isinstance(tensor, torch.FloatTensor)tensor = torch.randn(3, 5)assert isinstance(tensor, torch.FloatTensor)

使用torch.type()函數

type(new_type=None, async=False)

import torchtensor = torch.randn(3, 5)print(tensor)int_tensor = tensor.type(torch.IntTensor)print(int_tensor)

-0.4449 0.0332 0.5187 0.1271 2.2303 1.3961 -0.1542 0.8498 -0.3438 -0.2834-0.5554 0.1684 1.5216 2.4527 0.0379[torch.FloatTensor of size 3x5] 0 0 0 0 2 1 0 0 0 0 0 0 1 2 0[torch.IntTensor of size 3x5]

使用type_as(tesnor)將張量轉換為給定類型的張量

import torchtensor_1 = torch.FloatTensor(5)tensor_2 = torch.IntTensor([10, 20])tensor_1 = tensor_1.type_as(tensor_2)assert isinstance(tensor_1, torch.IntTensor)