mirror of
https://github.com/wassname/pytorch-for-numpy-users.git
synced 2026-06-27 16:10:21 +08:00
Mike Clark
149 lines
5.8 KiB
Markdown
149 lines
5.8 KiB
Markdown
# PyTorch for Numpy users.
|
|
|
|
[](https://travis-ci.com/wkentaro/pytorch-for-numpy-users)
|
|
|
|
[PyTorch](https://github.com/pytorch/pytorch.git) version of [_Torch for Numpy users_](https://github.com/torch/torch7/wiki/Torch-for-Numpy-users).
|
|
|
|
## Types
|
|
|
|
| Numpy | PyTorch |
|
|
|:-------------|:---------------------|
|
|
| `np.ndarray` | `torch.Tensor` |
|
|
| `np.float32` | `torch.FloatTensor` |
|
|
| `np.float64` | `torch.DoubleTensor` |
|
|
| `np.int8` | `torch.CharTensor` |
|
|
| `np.uint8` | `torch.ByteTensor` |
|
|
| `np.int16` | `torch.ShortTensor` |
|
|
| `np.int32` | `torch.IntTensor` |
|
|
| `np.int64` | `torch.LongTensor` |
|
|
|
|
|
|
## Constructors
|
|
|
|
### Ones and zeros
|
|
|
|
| Numpy | PyTorch |
|
|
|:-------------------|:---------------------------------------|
|
|
| `np.empty((2, 3))` | `torch.Tensor(2, 3)` |
|
|
| `np.empty_like(x)` | `x.new(x.size()).type(x.type())` |
|
|
| `np.eye` | `torch.eye` |
|
|
| `np.identity` | `torch.eye` |
|
|
| `np.ones` | `torch.ones` |
|
|
| `np.ones_like` | `torch.ones(x.size()).type(x.type())` |
|
|
| `np.zeros` | `torch.zeros` |
|
|
| `np.zeros_like` | `torch.zeros(x.size()).type(x.type())` |
|
|
|
|
### From existing data
|
|
|
|
| Numpy | PyTorch |
|
|
|:-----------------------------|:------------------------------------|
|
|
| `np.array([[1, 2], [3, 4]])` | `torch.Tensor([[1, 2], [3, 4])` |
|
|
| `x.copy()` | `x.clone()` |
|
|
| `np.fromfile(file)` | `torch.Tensor(torch.Storage(file))` |
|
|
| `np.frombuffer` | |
|
|
| `np.fromfunction` | |
|
|
| `np.fromiter` | |
|
|
| `np.fromstring` | |
|
|
| `np.loadtxt` | |
|
|
| `np.concatenate` | `torch.cat` |
|
|
|
|
### Numerical ranges
|
|
|
|
| Numpy | PyTorch |
|
|
|:-----------------------|:--------------------------|
|
|
| `np.arange(10)` | `torch.range(0, 9)` |
|
|
| `np.arange(2, 3, 0.1)` | `torch.range(2, 2.9, 10)` |
|
|
| `np.linspace` | `torch.linspace` |
|
|
| `np.logspace` | `torch.logspace` |
|
|
|
|
### Building matrices
|
|
|
|
| Numpy | PyTorch |
|
|
|:----------|:-------------|
|
|
| `np.diag` | `torch.diag` |
|
|
| `np.tril` | `torch.tril` |
|
|
| `np.triu` | `torch.triu` |
|
|
|
|
### Attributes
|
|
|
|
| Numpy | PyTorch |
|
|
|:------------|:---------------|
|
|
| `x.shape` | `x.size()` |
|
|
| `x.strides` | `x.stride()` |
|
|
| `x.ndim` | `x.dim()` |
|
|
| `x.data` | `x.data()` |
|
|
| `x.size` | `x.nelement()` |
|
|
| `x.dtype` | `x.type()` |
|
|
|
|
### Indexing
|
|
|
|
| Numpy | PyTorch |
|
|
|:----------------------|:-------------------------------|
|
|
| `x[0]` | `x[0]` |
|
|
| `x[:, 0]` | `x[:, 0]` |
|
|
| `x[indices]` | `x[torch.LongTensor(indices)]` |
|
|
| `np.take(x, indices)` | `x[torch.LongTensor(indices)]` |
|
|
| `x[x != 0]` | `x[x != 0]` |
|
|
| `x[::-1]` | [workaround exists](https://github.com/pytorch/pytorch/issues/229) |
|
|
|
|
### Shape manipulation
|
|
|
|
| Numpy | PyTorch |
|
|
|:-------------------|:-----------------|
|
|
| `x.reshape` | `x.view` |
|
|
| `x.resize` | `x.resize_` |
|
|
| | `x.resize_as_` |
|
|
| `x.transpose` | `x.permute` |
|
|
| `x.flatten()` | `x.view(-1)` |
|
|
| `x.squeeze` | `x.squeeze` |
|
|
| `x[:, np.newaxis]` | `x.unsqueeze(1)` |
|
|
|
|
### Item selection and manipulation
|
|
|
|
| Numpy | PyTorch |
|
|
|:-------------|:-----------------------------------------|
|
|
| `np.put` | |
|
|
| `x.repeat` | |
|
|
| `x.tile` | `x.repeat` |
|
|
| `np.choose` | |
|
|
| `np.sort` | `sorted, indices = torch.sort(x, [dim])` |
|
|
| `np.argsort` | `sorted, indices = torch.sort(x, [dim])` |
|
|
| `np.nonzero` | `torch.nonzero` |
|
|
| `np.where` | `torch.nonzero` |
|
|
|
|
### Calculation
|
|
|
|
| Numpy | PyTorch |
|
|
|:--------------|:--------------------------------------|
|
|
| `x.min` | `mins, indices = torch.min(x, [dim])` |
|
|
| `x.argmin` | `mins, indices = torch.min(x, [dim])` |
|
|
| `x.max` | `maxs, indices = torch.max(x, [dim])` |
|
|
| `x.argmax` | `maxs, indices = torch.max(x, [dim])` |
|
|
| `x.clip` | `x.clamp` |
|
|
| `x.round` | `x.round` |
|
|
| `np.floor(x)` | `x.floor()` |
|
|
| `np.ceil(x)` | `x.ceil()` |
|
|
| `x.trace` | `x.trace` |
|
|
| `x.sum` | `x.sum` |
|
|
| `x.cumsum` | `x.cumsum` |
|
|
| `x.mean` | `x.mean` |
|
|
| `x.std` | `x.std` |
|
|
| `x.prod` | `x.prod` |
|
|
| `x.cumprod` | `x.cumprod` |
|
|
| `x.all` | `(x == 1).sum() == x.nelement()` |
|
|
| `x.any` | `(x == 1).sum() > 0` |
|
|
|
|
### Arithmetic and comparison operations
|
|
|
|
| Numpy | PyTorch |
|
|
|:--------|:----------|
|
|
| `x.lt` | `x.lt` |
|
|
| `x.le` | `x.le` |
|
|
| `x.gt` | `x.gt` |
|
|
| `x.ge` | `x.ge` |
|
|
| `x.eq` | `x.eq` |
|
|
| `x.ne` | `x.ne` |
|
|
|
|
|
|
|