pytorch包学习心得--torch
1.torch.tensor(data, dtype=None, device=None, requires_grad=False) → Tensor
torch.tensor() always copies data. If you have a Tensor data and want to avoid a copy, use torch.Tensor.requires_grad_() or torch.Tensor.detach(). If you have a NumPy ndarrayand want to avoid a copy, use torch.from_numpy().
torch.tensor函数总会拷贝数据，除非使用其他函数可以避免拷贝。
2.torch.sparse_coo_tensor(indices, values, size=None, dtype=None, device=None, requires_grad=False) → Tensor
使用coordinate format matrix来进行稀疏矩阵的数据的存储。具体稀疏矩阵的表示方法可参考https://en.wikipedia.org/wiki/Sparse_matrix，其中，
一个稀疏矩阵：
能够被表示为如下一个三行的矩阵：
A = [ 5 8 3 6 ]
IA = [ 0 0 2 3 4 ]
JA = [ 0 1 2 1 ]A用来表示矩阵中的元素（从左到右，从上到下），IA[i]表示从第一行到第i行中的所有元素，IA[0]始终为0，而JA则表示相应的
元素在哪一列。但在这里indices用一个二维矩阵表示每个稀疏项的位置，第一维是行，第二维是列。
3.torch.as_tensor(data, dtype=None, device=None) → Tensor
转化一个ndarray成为tensor,如果同为cpu存储，则无需拷贝，否则需要拷贝，例如从cpu类型变为GPU类型，也就是赋值device相关属性。
4.torch.zeros_like(input, dtype=None, layout=None, device=None, requires_grad=False) → Tensor
其实在调用torch.zeros(input.size(),dtype=None, layout=None, device=None, requires_grad=False)
torch.ones_like()类似用法。
5.torch.arange(start=0, end, step=1, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) → Tensor 类似numpy.arange 函数，支持int和float，而python range函数不支持float类型。
6.torch.range(start=0, end, step=1, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) → Tensor 同上arange函数，与python自带的不一样，其支持float类型。
7.torch.logspace(start, end, steps=100, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) → Tensor
>>> torch.logspace(start=-10, end=10, steps=5)
tensor([ 1.0000e-10, 1.0000e-05, 1.0000e+00, 1.0000e+05, 1.0000e+10])
>>> torch.logspace(start=0.1, end=1.0, steps=5)
tensor([ 1.2589, 2.1135, 3.5481, 5.9566, 10.0000])这里start和end都是指数。
8.torch.eye(n, m=None, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) → Tensor
返回一个2维单位矩阵，
>>> torch.eye(3)
tensor([[ 1., 0., 0.],
[ 0., 1., 0.],
[ 0., 0., 1.]])9.torch.empty(*sizes, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) → Tensor
Returns a tensor filled with uninitialized data. The shape of the tensor is defined by the variable argument sizes.
10.torch.empty_like(input, dtype=None, layout=None, device=None, requires_grad=False) → Tensor
11.torch.full(size, fill_value, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) → Tensor
Returns a tensor of size size filled with fill_value.
感觉这个函数可以替代torch.zeros()和torch.ones().
12.torch.full_like(input, fill_value, out=None, dtype=None, layout=torch.strided, device=None, requires_grad=False) → Tensor 类似其它_like
13.torch.cat(tensors, dim=0, out=None) → Tensor
Concatenates the given sequence of seq tensors in the given dimension. All tensors must either have the same shape (except in the concatenating dimension) or be empty.
>>> x = torch.randn(2, 3)
>>> x
tensor([[ 0.6580, -1.0969, -0.4614],
[-0.1034, -0.5790, 0.1497]])
>>> torch.cat((x, x, x), 0)
tensor([[ 0.6580, -1.0969, -0.4614],
[-0.1034, -0.5790, 0.1497],
[ 0.6580, -1.0969, -0.4614],
[-0.1034, -0.5790, 0.1497],
[ 0.6580, -1.0969, -0.4614],
[-0.1034, -0.5790, 0.1497]])
>>> torch.cat((x, x, x), 1)
tensor([[ 0.6580, -1.0969, -0.4614, 0.6580, -1.0969, -0.4614, 0.6580,
-1.0969, -0.4614],
[-0.1034, -0.5790, 0.1497, -0.1034, -0.5790, 0.1497, -0.1034,
-0.5790, 0.1497]])14.torch.chunk(tensor, chunks, dim=0) → List of Tensors
Splits a tensor into a specific number of chunks.
Last chunk will be smaller if the tensor size along the given dimension dim is not divisible by chunks.
| Parameters: |
|
|---|
chunk函数用于分割torch数据。
15.torch.gather(input, dim, index, out=None) → Tensor
Gathers values along an axis specified by dim.
就是使用input中的数进行替换，规则如下：
out[i][j][k] = input[index[i][j][k]][j][k] # if dim == 0
out[i][j][k] = input[i][index[i][j][k]][k] # if dim == 1
out[i][j][k] = input[i][j][index[i][j][k]] # if dim == 2eg:
>>> t = torch.tensor([[1,2],[3,4]])
>>> g=torch.tensor([[0,0],[1,0]])
>>> torch.gather(t, 1, g)
tensor([[ 1, 1],
[ 4, 3]])这里是：[[t[0][g[0][0]],t[0][g[0][1]]],[t[1][g[1][0]],t[1][g[1][1]]]]
16.torch.index_select(input, dim, index, out=None) → Tensor
选择若干行和列
如下：
>>> x = torch.randn(3, 4)
>>> x
tensor([[ 0.1427, 0.0231, -0.5414, -1.0009],
[-0.4664, 0.2647, -0.1228, -1.1068],
[-1.1734, -0.6571, 0.7230, -0.6004]])
>>> indices = torch.tensor([0, 2])
>>> torch.index_select(x, 0, indices)
tensor([[ 0.1427, 0.0231, -0.5414, -1.0009],
[-1.1734, -0.6571, 0.7230, -0.6004]])
>>> torch.index_select(x, 1, indices)
tensor([[ 0.1427, -0.5414],
[-0.4664, -0.1228],
[-1.1734, 0.7230]])
17.torch.masked_select(input, mask, out=None) → Tensor
Returns a new 1-D tensor which indexes the input tensor according to the binary mask mask which is a ByteTensor.
The shapes of the mask tensor and the input tensor don’t need to match, but they must be broadcastable.
注意:返回的tensor是个拷贝而不是引用.
如下所示:
>>> x = torch.randn(3, 4)
>>> x
tensor([[ 0.3552, -2.3825, -0.8297, 0.3477],
[-1.2035, 1.2252, 0.5002, 0.6248],
[ 0.1307, -2.0608, 0.1244, 2.0139]])
>>> mask = x.ge(0.5)
>>> mask
tensor([[ 0, 0, 0, 0],
[ 0, 1, 1, 1],
[ 0, 0, 0, 1]], dtype=torch.uint8)
>>> torch.masked_select(x, mask)
tensor([ 1.2252, 0.5002, 0.6248, 2.0139])18.torch.narrow(input, dimension, start, length) → Tensor
Returns a new tensor that is a narrowed version of input tensor. The dimension dim is input from start to start + length. The returned tensor and input tensor share the same underlying storage.
选择一维tensor进行切片操作.
19.torch.nonzero(input, out=None) → LongTensor
Returns a tensor containing the indices of all non-zero elements of input. Each row in the result contains the indices of a non-zero element in input.
返回非0点的点的坐标:
>>> torch.nonzero(torch.tensor([1, 1, 1, 0, 1]))
tensor([[ 0],
[ 1],
[ 2],
[ 4]])
>>> torch.nonzero(torch.tensor([[0.6, 0.0, 0.0, 0.0],
[0.0, 0.4, 0.0, 0.0],
[0.0, 0.0, 1.2, 0.0],
[0.0, 0.0, 0.0,-0.4]]))
tensor([[ 0, 0],
[ 1, 1],
[ 2, 2],
[ 3, 3]])20.torch.reshape(input, shape) → Tensor
Returns a tensor with the same data and number of elements as input, but with the specified shape. When possible, the returned tensor will be a view of input. Otherwise, it will be a copy. Contiguous inputs and inputs with compatible strides can be reshaped without copying, but you should not depend on the copying vs. viewing behavior.
torch.Tensor.view() 也有类似功能.
A single dimension may be -1, in which case it’s inferred from the remaining dimensions and the number of elements in input.
21.torch.split(tensor, split_size_or_sections, dim=0)
Splits the tensor into chunks.
和torch.chunk一样.
22.torch.squeeze(input, dim=None, out=None) → Tensor
Returns a tensor with all the dimensions of input of size 1 removed.
把维度为1的删除掉.
>>> x = torch.zeros(2, 1, 2, 1, 2)
>>> x.size()
torch.Size([2, 1, 2, 1, 2])
>>> y = torch.squeeze(x)
>>> y.size()
torch.Size([2, 2, 2])
>>> y = torch.squeeze(x, 0)
>>> y.size()
torch.Size([2, 1, 2, 1, 2])
>>> y = torch.squeeze(x, 1)
>>> y.size()
torch.Size([2, 2, 1, 2])23.orch.stack(seq, dim=0, out=None) → Tensor
Concatenates sequence of tensors along a new dimension.
All tensors need to be of the same size.
类似torch.cat()
24.torch.t(input) → Tensor
Expects input to be a matrix (2-D tensor) and transposes dimensions 0 and 1.
和transpose(input, 0, 1)效果一样
转置
25.torch.transpose(input, dim0, dim1) → Tensor
| Parameters: |
|
|---|
26.torch.take(input, indices) → Tensor
提取数值用,返回的是个一维的tensor,就是把input展开为一维,然后取数.
>>> src = torch.tensor([[4, 3, 5],
[6, 7, 8]])
>>> torch.take(src, torch.tensor([0, 2, 5]))
tensor([ 4, 5, 8])27.torch.unbind(tensor, dim=0) → seq
Removes a tensor dimension.类似split.
>>> torch.unbind(torch.tensor([[1, 2, 3],
>>> [4, 5, 6],
>>> [7, 8, 9]]))
(tensor([1, 2, 3]), tensor([4, 5, 6]), tensor([7, 8, 9]))28.
torch.where(condition, x, y) → Tensor
Return a tensor of elements selected from either x or y, depending on condition.
>>> x = torch.randn(3, 2)
>>> y = torch.ones(3, 2)
>>> x
tensor([[-0.4620, 0.3139],
[ 0.3898, -0.7197],
[ 0.0478, -0.1657]])
>>> torch.where(x > 0, x, y)
tensor([[ 1.0000, 0.3139],
[ 0.3898, 1.0000],
[ 0.0478, 1.0000]])29.torch.bernoulli(input, *, generator=None, out=None) → Tensor
生成伯努利分布,input里面都是大于等于0小于等于1的数,这些都是概率,根据概率生成模型.
>>> a = torch.empty(3, 3).uniform_(0, 1) # generate a uniform random matrix with range [0, 1]
>>> a
tensor([[ 0.1737, 0.0950, 0.3609],
[ 0.7148, 0.0289, 0.2676],
[ 0.9456, 0.8937, 0.7202]])
>>> torch.bernoulli(a)
tensor([[ 1., 0., 0.],
[ 0., 0., 0.],
[ 1., 1., 1.]])
>>> a = torch.ones(3, 3) # probability of drawing "1" is 1
>>> torch.bernoulli(a)
tensor([[ 1., 1., 1.],
[ 1., 1., 1.],
[ 1., 1., 1.]])
>>> a = torch.zeros(3, 3) # probability of drawing "1" is 0
>>> torch.bernoulli(a)
tensor([[ 0., 0., 0.],
[ 0., 0., 0.],
[ 0., 0., 0.]])30.torch.save(obj, f, pickle_module=
Saves an object to a disk file.
保存数据
>>> # Save to file >>> x = torch.tensor([0, 1, 2, 3, 4]) >>> torch.save(x, 'tensor.pt') >>> # Save to io.BytesIO buffer >>> buffer = io.BytesIO() >>> torch.save(x, buffer)
31.torch.load(f, map_location=None, pickle_module=
Loads an object saved with torch.save() from a file.
>>> torch.load('tensors.pt')
# Load all tensors onto the CPU
>>> torch.load('tensors.pt', map_location=torch.device('cpu'))
# Load all tensors onto the CPU, using a function
>>> torch.load('tensors.pt', map_location=lambda storage, loc: storage)
# Load all tensors onto GPU 1
>>> torch.load('tensors.pt', map_location=lambda storage, loc: storage.cuda(1))
# Map tensors from GPU 1 to GPU 0
>>> torch.load('tensors.pt', map_location={'cuda:1':'cuda:0'})
# Load tensor from io.BytesIO object
>>> with open('tensor.pt') as f:
buffer = io.BytesIO(f.read())
>>> torch.load(buffer)