Tenseur for torch and numpy users¶
Tenseur |
Torch |
Numpy |
|---|---|---|
ten::tensor<T, Rank> |
torch.Tensor |
np.ndarray |
ten::matrix<T> |
torch.Tensor |
np.ndarray |
ten::vector<T> |
torch.Tensor |
np.ndarray |
ten::stensor<Dims…> |
x |
x |
ten::smatrix<Rows, Cols> |
x |
x |
ten::svector<Size> |
x |
x |
float |
torch.float32; torch.float |
np.float32 |
double |
torch.float64; torch.double |
np.float64 |
x |
torch.float16; torch.half |
np.float16 |
x |
torch.int8 |
np.int8 |
x |
torch.uint8 |
np.uint8 |
x |
torch.int16; torch.short |
np.int16 |
int32_t |
torch.int32; torch.int |
np.int32 |
uint32_t |
x |
x |
int64_t |
torch.int64; torch.long |
np.int64 |
uint64_t |
x |
x |