types.backends.Device

types.backends.Device

Device identifier for hardware acceleration.

Common values: - ‘cpu’: CPU computation - ‘cuda’: NVIDIA GPU (any available) - ‘cuda:0’, ‘cuda:1’, …: Specific NVIDIA GPU - ‘mps’: Apple Metal (M1/M2/M3 Macs) - ‘tpu’: Google TPU (JAX only)

Examples

>>> device: Device = 'cuda:0'
>>> system.set_preferred_device('cpu')
>>> 
>>> # PyTorch usage
>>> import torch
>>> tensor = torch.tensor(data, device='cuda')
>>> 
>>> # JAX usage (via jax.devices)
>>> import jax
>>> device = jax.devices('gpu')[0]

# types.backends.Device { #cdesym.types.backends.Device }

`types.backends.Device`

Device identifier for hardware acceleration.

Common values:
- 'cpu': CPU computation
- 'cuda': NVIDIA GPU (any available)
- 'cuda:0', 'cuda:1', ...: Specific NVIDIA GPU
- 'mps': Apple Metal (M1/M2/M3 Macs)
- 'tpu': Google TPU (JAX only)

## Examples {.doc-section .doc-section-examples}

```python
>>> device: Device = 'cuda:0'
>>> system.set_preferred_device('cpu')
>>> 
>>> # PyTorch usage
>>> import torch
>>> tensor = torch.tensor(data, device='cuda')
>>> 
>>> # JAX usage (via jax.devices)
>>> import jax
>>> device = jax.devices('gpu')[0]
```