Tensor Operations¶
Core tensor operations and data structures in Genesis.
Overview¶
Genesis tensors provide the fundamental data structure for all computations, similar to PyTorch tensors but optimized for our dual backend architecture.
Creating Tensors¶
Python
import genesis
# Create tensors
x = genesis.tensor([1, 2, 3, 4])
y = genesis.zeros(3, 4)
z = genesis.randn(2, 3, device='cuda')
Tensor Operations¶
Device Management¶
Python
# Move tensors between devices
cpu_tensor = gpu_tensor.cpu()
gpu_tensor = cpu_tensor.cuda()
gpu_tensor = cpu_tensor.to('cuda')
This documentation is under construction. More detailed tensor API documentation will be added.
See Also¶
- Autograd - Automatic differentiation with tensors
- CUDA Storage - GPU memory management